1 /*
2 * Copyright (c) 2001, 2014, 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 "gc_implementation/shared/gcTimer.hpp"
40 #include "gc_implementation/shared/gcTrace.hpp"
41 #include "gc_implementation/shared/gcTraceTime.hpp"
42 #include "memory/genOopClosures.inline.hpp"
43 #include "memory/referencePolicy.hpp"
44 #include "memory/resourceArea.hpp"
45 #include "oops/oop.inline.hpp"
46 #include "runtime/handles.inline.hpp"
47 #include "runtime/java.hpp"
48 #include "services/memTracker.hpp"
49
50 // Concurrent marking bit map wrapper
51
52 CMBitMapRO::CMBitMapRO(int shifter) :
53 _bm(),
54 _shifter(shifter) {
55 _bmStartWord = 0;
56 _bmWordSize = 0;
57 }
58
59 HeapWord* CMBitMapRO::getNextMarkedWordAddress(HeapWord* addr,
60 HeapWord* limit) const {
61 // First we must round addr *up* to a possible object boundary.
62 addr = (HeapWord*)align_size_up((intptr_t)addr,
63 HeapWordSize << _shifter);
64 size_t addrOffset = heapWordToOffset(addr);
65 if (limit == NULL) {
66 limit = _bmStartWord + _bmWordSize;
67 }
68 size_t limitOffset = heapWordToOffset(limit);
69 size_t nextOffset = _bm.get_next_one_offset(addrOffset, limitOffset);
70 HeapWord* nextAddr = offsetToHeapWord(nextOffset);
71 assert(nextAddr >= addr, "get_next_one postcondition");
72 assert(nextAddr == limit || isMarked(nextAddr),
73 "get_next_one postcondition");
74 return nextAddr;
75 }
76
77 HeapWord* CMBitMapRO::getNextUnmarkedWordAddress(HeapWord* addr,
78 HeapWord* limit) const {
79 size_t addrOffset = heapWordToOffset(addr);
80 if (limit == NULL) {
81 limit = _bmStartWord + _bmWordSize;
82 }
83 size_t limitOffset = heapWordToOffset(limit);
84 size_t nextOffset = _bm.get_next_zero_offset(addrOffset, limitOffset);
85 HeapWord* nextAddr = offsetToHeapWord(nextOffset);
86 assert(nextAddr >= addr, "get_next_one postcondition");
87 assert(nextAddr == limit || !isMarked(nextAddr),
88 "get_next_one postcondition");
89 return nextAddr;
90 }
91
92 int CMBitMapRO::heapWordDiffToOffsetDiff(size_t diff) const {
93 assert((diff & ((1 << _shifter) - 1)) == 0, "argument check");
94 return (int) (diff >> _shifter);
95 }
96
97 #ifndef PRODUCT
98 bool CMBitMapRO::covers(ReservedSpace heap_rs) const {
99 // assert(_bm.map() == _virtual_space.low(), "map inconsistency");
100 assert(((size_t)_bm.size() * ((size_t)1 << _shifter)) == _bmWordSize,
101 "size inconsistency");
102 return _bmStartWord == (HeapWord*)(heap_rs.base()) &&
103 _bmWordSize == heap_rs.size()>>LogHeapWordSize;
104 }
105 #endif
106
107 void CMBitMapRO::print_on_error(outputStream* st, const char* prefix) const {
108 _bm.print_on_error(st, prefix);
109 }
110
111 bool CMBitMap::allocate(ReservedSpace heap_rs) {
112 _bmStartWord = (HeapWord*)(heap_rs.base());
113 _bmWordSize = heap_rs.size()/HeapWordSize; // heap_rs.size() is in bytes
114 ReservedSpace brs(ReservedSpace::allocation_align_size_up(
115 (_bmWordSize >> (_shifter + LogBitsPerByte)) + 1));
116 if (!brs.is_reserved()) {
117 warning("ConcurrentMark marking bit map allocation failure");
118 return false;
119 }
120 MemTracker::record_virtual_memory_type((address)brs.base(), mtGC);
121 // For now we'll just commit all of the bit map up front.
122 // Later on we'll try to be more parsimonious with swap.
123 if (!_virtual_space.initialize(brs, brs.size())) {
124 warning("ConcurrentMark marking bit map backing store failure");
125 return false;
126 }
127 assert(_virtual_space.committed_size() == brs.size(),
128 "didn't reserve backing store for all of concurrent marking bit map?");
129 _bm.set_map((uintptr_t*)_virtual_space.low());
130 assert(_virtual_space.committed_size() << (_shifter + LogBitsPerByte) >=
131 _bmWordSize, "inconsistency in bit map sizing");
132 _bm.set_size(_bmWordSize >> _shifter);
133 return true;
134 }
135
136 void CMBitMap::clearAll() {
137 _bm.clear();
138 return;
139 }
140
141 void CMBitMap::markRange(MemRegion mr) {
142 mr.intersection(MemRegion(_bmStartWord, _bmWordSize));
143 assert(!mr.is_empty(), "unexpected empty region");
144 assert((offsetToHeapWord(heapWordToOffset(mr.end())) ==
145 ((HeapWord *) mr.end())),
146 "markRange memory region end is not card aligned");
147 // convert address range into offset range
148 _bm.at_put_range(heapWordToOffset(mr.start()),
149 heapWordToOffset(mr.end()), true);
150 }
151
152 void CMBitMap::clearRange(MemRegion mr) {
153 mr.intersection(MemRegion(_bmStartWord, _bmWordSize));
154 assert(!mr.is_empty(), "unexpected empty region");
155 // convert address range into offset range
156 _bm.at_put_range(heapWordToOffset(mr.start()),
157 heapWordToOffset(mr.end()), false);
158 }
159
160 MemRegion CMBitMap::getAndClearMarkedRegion(HeapWord* addr,
161 HeapWord* end_addr) {
162 HeapWord* start = getNextMarkedWordAddress(addr);
163 start = MIN2(start, end_addr);
164 HeapWord* end = getNextUnmarkedWordAddress(start);
165 end = MIN2(end, end_addr);
166 assert(start <= end, "Consistency check");
167 MemRegion mr(start, end);
168 if (!mr.is_empty()) {
169 clearRange(mr);
170 }
171 return mr;
172 }
173
174 CMMarkStack::CMMarkStack(ConcurrentMark* cm) :
175 _base(NULL), _cm(cm)
176 #ifdef ASSERT
177 , _drain_in_progress(false)
178 , _drain_in_progress_yields(false)
179 #endif
180 {}
181
182 bool CMMarkStack::allocate(size_t capacity) {
183 // allocate a stack of the requisite depth
184 ReservedSpace rs(ReservedSpace::allocation_align_size_up(capacity * sizeof(oop)));
185 if (!rs.is_reserved()) {
186 warning("ConcurrentMark MarkStack allocation failure");
187 return false;
188 }
189 MemTracker::record_virtual_memory_type((address)rs.base(), mtGC);
190 if (!_virtual_space.initialize(rs, rs.size())) {
191 warning("ConcurrentMark MarkStack backing store failure");
192 // Release the virtual memory reserved for the marking stack
193 rs.release();
194 return false;
195 }
196 assert(_virtual_space.committed_size() == rs.size(),
197 "Didn't reserve backing store for all of ConcurrentMark stack?");
198 _base = (oop*) _virtual_space.low();
199 setEmpty();
200 _capacity = (jint) capacity;
201 _saved_index = -1;
202 _should_expand = false;
203 NOT_PRODUCT(_max_depth = 0);
204 return true;
205 }
206
207 void CMMarkStack::expand() {
208 // Called, during remark, if we've overflown the marking stack during marking.
209 assert(isEmpty(), "stack should been emptied while handling overflow");
210 assert(_capacity <= (jint) MarkStackSizeMax, "stack bigger than permitted");
211 // Clear expansion flag
212 _should_expand = false;
213 if (_capacity == (jint) MarkStackSizeMax) {
214 if (PrintGCDetails && Verbose) {
215 gclog_or_tty->print_cr(" (benign) Can't expand marking stack capacity, at max size limit");
216 }
217 return;
218 }
219 // Double capacity if possible
220 jint new_capacity = MIN2(_capacity*2, (jint) MarkStackSizeMax);
221 // Do not give up existing stack until we have managed to
222 // get the double capacity that we desired.
223 ReservedSpace rs(ReservedSpace::allocation_align_size_up(new_capacity *
224 sizeof(oop)));
225 if (rs.is_reserved()) {
226 // Release the backing store associated with old stack
227 _virtual_space.release();
228 // Reinitialize virtual space for new stack
229 if (!_virtual_space.initialize(rs, rs.size())) {
230 fatal("Not enough swap for expanded marking stack capacity");
231 }
232 _base = (oop*)(_virtual_space.low());
233 _index = 0;
234 _capacity = new_capacity;
235 } else {
236 if (PrintGCDetails && Verbose) {
237 // Failed to double capacity, continue;
238 gclog_or_tty->print(" (benign) Failed to expand marking stack capacity from "
239 SIZE_FORMAT"K to " SIZE_FORMAT"K",
240 _capacity / K, new_capacity / K);
241 }
242 }
243 }
244
245 void CMMarkStack::set_should_expand() {
246 // If we're resetting the marking state because of an
247 // marking stack overflow, record that we should, if
248 // possible, expand the stack.
249 _should_expand = _cm->has_overflown();
250 }
251
252 CMMarkStack::~CMMarkStack() {
253 if (_base != NULL) {
254 _base = NULL;
255 _virtual_space.release();
256 }
257 }
258
259 void CMMarkStack::par_push(oop ptr) {
260 while (true) {
261 if (isFull()) {
262 _overflow = true;
263 return;
264 }
265 // Otherwise...
266 jint index = _index;
267 jint next_index = index+1;
268 jint res = Atomic::cmpxchg(next_index, &_index, index);
269 if (res == index) {
270 _base[index] = ptr;
271 // Note that we don't maintain this atomically. We could, but it
272 // doesn't seem necessary.
273 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
274 return;
275 }
276 // Otherwise, we need to try again.
277 }
278 }
279
280 void CMMarkStack::par_adjoin_arr(oop* ptr_arr, int n) {
281 while (true) {
282 if (isFull()) {
283 _overflow = true;
284 return;
285 }
286 // Otherwise...
287 jint index = _index;
288 jint next_index = index + n;
289 if (next_index > _capacity) {
290 _overflow = true;
291 return;
292 }
293 jint res = Atomic::cmpxchg(next_index, &_index, index);
294 if (res == index) {
295 for (int i = 0; i < n; i++) {
296 int ind = index + i;
297 assert(ind < _capacity, "By overflow test above.");
298 _base[ind] = ptr_arr[i];
299 }
300 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
301 return;
302 }
303 // Otherwise, we need to try again.
304 }
305 }
306
307 void CMMarkStack::par_push_arr(oop* ptr_arr, int n) {
308 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
309 jint start = _index;
310 jint next_index = start + n;
311 if (next_index > _capacity) {
312 _overflow = true;
313 return;
314 }
315 // Otherwise.
316 _index = next_index;
317 for (int i = 0; i < n; i++) {
318 int ind = start + i;
319 assert(ind < _capacity, "By overflow test above.");
320 _base[ind] = ptr_arr[i];
321 }
322 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
323 }
324
325 bool CMMarkStack::par_pop_arr(oop* ptr_arr, int max, int* n) {
326 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
327 jint index = _index;
328 if (index == 0) {
329 *n = 0;
330 return false;
331 } else {
332 int k = MIN2(max, index);
333 jint new_ind = index - k;
334 for (int j = 0; j < k; j++) {
335 ptr_arr[j] = _base[new_ind + j];
336 }
337 _index = new_ind;
338 *n = k;
339 return true;
340 }
341 }
342
343 template<class OopClosureClass>
344 bool CMMarkStack::drain(OopClosureClass* cl, CMBitMap* bm, bool yield_after) {
345 assert(!_drain_in_progress || !_drain_in_progress_yields || yield_after
346 || SafepointSynchronize::is_at_safepoint(),
347 "Drain recursion must be yield-safe.");
348 bool res = true;
349 debug_only(_drain_in_progress = true);
350 debug_only(_drain_in_progress_yields = yield_after);
351 while (!isEmpty()) {
352 oop newOop = pop();
353 assert(G1CollectedHeap::heap()->is_in_reserved(newOop), "Bad pop");
354 assert(newOop->is_oop(), "Expected an oop");
355 assert(bm == NULL || bm->isMarked((HeapWord*)newOop),
356 "only grey objects on this stack");
357 newOop->oop_iterate(cl);
358 if (yield_after && _cm->do_yield_check()) {
359 res = false;
360 break;
361 }
362 }
363 debug_only(_drain_in_progress = false);
364 return res;
365 }
366
367 void CMMarkStack::note_start_of_gc() {
368 assert(_saved_index == -1,
369 "note_start_of_gc()/end_of_gc() bracketed incorrectly");
370 _saved_index = _index;
371 }
372
373 void CMMarkStack::note_end_of_gc() {
374 // This is intentionally a guarantee, instead of an assert. If we
375 // accidentally add something to the mark stack during GC, it
376 // will be a correctness issue so it's better if we crash. we'll
377 // only check this once per GC anyway, so it won't be a performance
378 // issue in any way.
379 guarantee(_saved_index == _index,
380 err_msg("saved index: %d index: %d", _saved_index, _index));
381 _saved_index = -1;
382 }
383
384 void CMMarkStack::oops_do(OopClosure* f) {
385 assert(_saved_index == _index,
386 err_msg("saved index: %d index: %d", _saved_index, _index));
387 for (int i = 0; i < _index; i += 1) {
388 f->do_oop(&_base[i]);
389 }
390 }
391
392 bool ConcurrentMark::not_yet_marked(oop obj) const {
393 return _g1h->is_obj_ill(obj);
394 }
395
396 CMRootRegions::CMRootRegions() :
397 _young_list(NULL), _cm(NULL), _scan_in_progress(false),
398 _should_abort(false), _next_survivor(NULL) { }
399
400 void CMRootRegions::init(G1CollectedHeap* g1h, ConcurrentMark* cm) {
401 _young_list = g1h->young_list();
402 _cm = cm;
403 }
404
405 void CMRootRegions::prepare_for_scan() {
406 assert(!scan_in_progress(), "pre-condition");
407
408 // Currently, only survivors can be root regions.
409 assert(_next_survivor == NULL, "pre-condition");
410 _next_survivor = _young_list->first_survivor_region();
411 _scan_in_progress = (_next_survivor != NULL);
412 _should_abort = false;
413 }
414
415 HeapRegion* CMRootRegions::claim_next() {
416 if (_should_abort) {
417 // If someone has set the should_abort flag, we return NULL to
418 // force the caller to bail out of their loop.
419 return NULL;
420 }
421
422 // Currently, only survivors can be root regions.
423 HeapRegion* res = _next_survivor;
424 if (res != NULL) {
425 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
426 // Read it again in case it changed while we were waiting for the lock.
427 res = _next_survivor;
428 if (res != NULL) {
429 if (res == _young_list->last_survivor_region()) {
430 // We just claimed the last survivor so store NULL to indicate
431 // that we're done.
432 _next_survivor = NULL;
433 } else {
434 _next_survivor = res->get_next_young_region();
435 }
436 } else {
437 // Someone else claimed the last survivor while we were trying
438 // to take the lock so nothing else to do.
439 }
440 }
441 assert(res == NULL || res->is_survivor(), "post-condition");
442
443 return res;
444 }
445
446 void CMRootRegions::scan_finished() {
447 assert(scan_in_progress(), "pre-condition");
448
449 // Currently, only survivors can be root regions.
450 if (!_should_abort) {
451 assert(_next_survivor == NULL, "we should have claimed all survivors");
452 }
453 _next_survivor = NULL;
454
455 {
456 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
457 _scan_in_progress = false;
458 RootRegionScan_lock->notify_all();
459 }
460 }
461
462 bool CMRootRegions::wait_until_scan_finished() {
463 if (!scan_in_progress()) return false;
464
465 {
466 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
467 while (scan_in_progress()) {
468 RootRegionScan_lock->wait(Mutex::_no_safepoint_check_flag);
469 }
470 }
471 return true;
472 }
473
474 #ifdef _MSC_VER // the use of 'this' below gets a warning, make it go away
475 #pragma warning( disable:4355 ) // 'this' : used in base member initializer list
476 #endif // _MSC_VER
477
478 uint ConcurrentMark::scale_parallel_threads(uint n_par_threads) {
479 return MAX2((n_par_threads + 2) / 4, 1U);
480 }
481
482 ConcurrentMark::ConcurrentMark(G1CollectedHeap* g1h, ReservedSpace heap_rs) :
483 _g1h(g1h),
484 _markBitMap1(log2_intptr(MinObjAlignment)),
485 _markBitMap2(log2_intptr(MinObjAlignment)),
486 _parallel_marking_threads(0),
487 _max_parallel_marking_threads(0),
488 _sleep_factor(0.0),
489 _marking_task_overhead(1.0),
490 _cleanup_sleep_factor(0.0),
491 _cleanup_task_overhead(1.0),
492 _cleanup_list("Cleanup List"),
493 _region_bm((BitMap::idx_t)(g1h->max_regions()), false /* in_resource_area*/),
494 _card_bm((heap_rs.size() + CardTableModRefBS::card_size - 1) >>
495 CardTableModRefBS::card_shift,
496 false /* in_resource_area*/),
497
498 _prevMarkBitMap(&_markBitMap1),
499 _nextMarkBitMap(&_markBitMap2),
500
501 _markStack(this),
502 // _finger set in set_non_marking_state
503
504 _max_worker_id(MAX2((uint)ParallelGCThreads, 1U)),
505 // _active_tasks set in set_non_marking_state
506 // _tasks set inside the constructor
507 _task_queues(new CMTaskQueueSet((int) _max_worker_id)),
508 _terminator(ParallelTaskTerminator((int) _max_worker_id, _task_queues)),
509
510 _has_overflown(false),
511 _concurrent(false),
512 _has_aborted(false),
513 _restart_for_overflow(false),
514 _concurrent_marking_in_progress(false),
515
516 // _verbose_level set below
517
518 _init_times(),
519 _remark_times(), _remark_mark_times(), _remark_weak_ref_times(),
520 _cleanup_times(),
521 _total_counting_time(0.0),
522 _total_rs_scrub_time(0.0),
523
524 _parallel_workers(NULL),
525
526 _count_card_bitmaps(NULL),
527 _count_marked_bytes(NULL),
528 _completed_initialization(false) {
529 CMVerboseLevel verbose_level = (CMVerboseLevel) G1MarkingVerboseLevel;
530 if (verbose_level < no_verbose) {
531 verbose_level = no_verbose;
532 }
533 if (verbose_level > high_verbose) {
534 verbose_level = high_verbose;
535 }
536 _verbose_level = verbose_level;
537
538 if (verbose_low()) {
539 gclog_or_tty->print_cr("[global] init, heap start = "PTR_FORMAT", "
540 "heap end = " PTR_FORMAT, p2i(_heap_start), p2i(_heap_end));
541 }
542
543 if (!_markBitMap1.allocate(heap_rs)) {
544 warning("Failed to allocate first CM bit map");
545 return;
546 }
547 if (!_markBitMap2.allocate(heap_rs)) {
548 warning("Failed to allocate second CM bit map");
549 return;
550 }
551
552 // Create & start a ConcurrentMark thread.
553 _cmThread = new ConcurrentMarkThread(this);
554 assert(cmThread() != NULL, "CM Thread should have been created");
555 assert(cmThread()->cm() != NULL, "CM Thread should refer to this cm");
556 if (_cmThread->osthread() == NULL) {
557 vm_shutdown_during_initialization("Could not create ConcurrentMarkThread");
558 }
559
560 assert(CGC_lock != NULL, "Where's the CGC_lock?");
561 assert(_markBitMap1.covers(heap_rs), "_markBitMap1 inconsistency");
562 assert(_markBitMap2.covers(heap_rs), "_markBitMap2 inconsistency");
563
564 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set();
565 satb_qs.set_buffer_size(G1SATBBufferSize);
566
567 _root_regions.init(_g1h, this);
568
569 if (ConcGCThreads > ParallelGCThreads) {
570 warning("Can't have more ConcGCThreads (" UINTX_FORMAT ") "
571 "than ParallelGCThreads (" UINTX_FORMAT ").",
572 ConcGCThreads, ParallelGCThreads);
573 return;
574 }
575 if (ParallelGCThreads == 0) {
576 // if we are not running with any parallel GC threads we will not
577 // spawn any marking threads either
578 _parallel_marking_threads = 0;
579 _max_parallel_marking_threads = 0;
580 _sleep_factor = 0.0;
581 _marking_task_overhead = 1.0;
582 } else {
583 if (!FLAG_IS_DEFAULT(ConcGCThreads) && ConcGCThreads > 0) {
584 // Note: ConcGCThreads has precedence over G1MarkingOverheadPercent
585 // if both are set
586 _sleep_factor = 0.0;
587 _marking_task_overhead = 1.0;
588 } else if (G1MarkingOverheadPercent > 0) {
589 // We will calculate the number of parallel marking threads based
590 // on a target overhead with respect to the soft real-time goal
591 double marking_overhead = (double) G1MarkingOverheadPercent / 100.0;
592 double overall_cm_overhead =
593 (double) MaxGCPauseMillis * marking_overhead /
594 (double) GCPauseIntervalMillis;
595 double cpu_ratio = 1.0 / (double) os::processor_count();
596 double marking_thread_num = ceil(overall_cm_overhead / cpu_ratio);
597 double marking_task_overhead =
598 overall_cm_overhead / marking_thread_num *
599 (double) os::processor_count();
600 double sleep_factor =
601 (1.0 - marking_task_overhead) / marking_task_overhead;
602
603 FLAG_SET_ERGO(uintx, ConcGCThreads, (uint) marking_thread_num);
604 _sleep_factor = sleep_factor;
605 _marking_task_overhead = marking_task_overhead;
606 } else {
607 // Calculate the number of parallel marking threads by scaling
608 // the number of parallel GC threads.
609 uint marking_thread_num = scale_parallel_threads((uint) ParallelGCThreads);
610 FLAG_SET_ERGO(uintx, ConcGCThreads, marking_thread_num);
611 _sleep_factor = 0.0;
612 _marking_task_overhead = 1.0;
613 }
614
615 assert(ConcGCThreads > 0, "Should have been set");
616 _parallel_marking_threads = (uint) ConcGCThreads;
617 _max_parallel_marking_threads = _parallel_marking_threads;
618
619 if (parallel_marking_threads() > 1) {
620 _cleanup_task_overhead = 1.0;
621 } else {
622 _cleanup_task_overhead = marking_task_overhead();
623 }
624 _cleanup_sleep_factor =
625 (1.0 - cleanup_task_overhead()) / cleanup_task_overhead();
626
627 #if 0
628 gclog_or_tty->print_cr("Marking Threads %d", parallel_marking_threads());
629 gclog_or_tty->print_cr("CM Marking Task Overhead %1.4lf", marking_task_overhead());
630 gclog_or_tty->print_cr("CM Sleep Factor %1.4lf", sleep_factor());
631 gclog_or_tty->print_cr("CL Marking Task Overhead %1.4lf", cleanup_task_overhead());
632 gclog_or_tty->print_cr("CL Sleep Factor %1.4lf", cleanup_sleep_factor());
633 #endif
634
635 guarantee(parallel_marking_threads() > 0, "peace of mind");
636 _parallel_workers = new FlexibleWorkGang("G1 Parallel Marking Threads",
637 _max_parallel_marking_threads, false, true);
638 if (_parallel_workers == NULL) {
639 vm_exit_during_initialization("Failed necessary allocation.");
640 } else {
641 _parallel_workers->initialize_workers();
642 }
643 }
644
645 if (FLAG_IS_DEFAULT(MarkStackSize)) {
646 uintx mark_stack_size =
647 MIN2(MarkStackSizeMax,
648 MAX2(MarkStackSize, (uintx) (parallel_marking_threads() * TASKQUEUE_SIZE)));
649 // Verify that the calculated value for MarkStackSize is in range.
650 // It would be nice to use the private utility routine from Arguments.
651 if (!(mark_stack_size >= 1 && mark_stack_size <= MarkStackSizeMax)) {
652 warning("Invalid value calculated for MarkStackSize (" UINTX_FORMAT "): "
653 "must be between " UINTX_FORMAT " and " UINTX_FORMAT,
654 mark_stack_size, (uintx) 1, MarkStackSizeMax);
655 return;
656 }
657 FLAG_SET_ERGO(uintx, MarkStackSize, mark_stack_size);
658 } else {
659 // Verify MarkStackSize is in range.
660 if (FLAG_IS_CMDLINE(MarkStackSize)) {
661 if (FLAG_IS_DEFAULT(MarkStackSizeMax)) {
662 if (!(MarkStackSize >= 1 && MarkStackSize <= MarkStackSizeMax)) {
663 warning("Invalid value specified for MarkStackSize (" UINTX_FORMAT "): "
664 "must be between " UINTX_FORMAT " and " UINTX_FORMAT,
665 MarkStackSize, (uintx) 1, MarkStackSizeMax);
666 return;
667 }
668 } else if (FLAG_IS_CMDLINE(MarkStackSizeMax)) {
669 if (!(MarkStackSize >= 1 && MarkStackSize <= MarkStackSizeMax)) {
670 warning("Invalid value specified for MarkStackSize (" UINTX_FORMAT ")"
671 " or for MarkStackSizeMax (" UINTX_FORMAT ")",
672 MarkStackSize, MarkStackSizeMax);
673 return;
674 }
675 }
676 }
677 }
678
679 if (!_markStack.allocate(MarkStackSize)) {
680 warning("Failed to allocate CM marking stack");
681 return;
682 }
683
684 _tasks = NEW_C_HEAP_ARRAY(CMTask*, _max_worker_id, mtGC);
685 _accum_task_vtime = NEW_C_HEAP_ARRAY(double, _max_worker_id, mtGC);
686
687 _count_card_bitmaps = NEW_C_HEAP_ARRAY(BitMap, _max_worker_id, mtGC);
688 _count_marked_bytes = NEW_C_HEAP_ARRAY(size_t*, _max_worker_id, mtGC);
689
690 BitMap::idx_t card_bm_size = _card_bm.size();
691
692 // so that the assertion in MarkingTaskQueue::task_queue doesn't fail
693 _active_tasks = _max_worker_id;
694
695 size_t max_regions = (size_t) _g1h->max_regions();
696 for (uint i = 0; i < _max_worker_id; ++i) {
697 CMTaskQueue* task_queue = new CMTaskQueue();
698 task_queue->initialize();
699 _task_queues->register_queue(i, task_queue);
700
701 _count_card_bitmaps[i] = BitMap(card_bm_size, false);
702 _count_marked_bytes[i] = NEW_C_HEAP_ARRAY(size_t, max_regions, mtGC);
703
704 _tasks[i] = new CMTask(i, this,
705 _count_marked_bytes[i],
706 &_count_card_bitmaps[i],
707 task_queue, _task_queues);
708
709 _accum_task_vtime[i] = 0.0;
710 }
711
712 // Calculate the card number for the bottom of the heap. Used
713 // in biasing indexes into the accounting card bitmaps.
714 _heap_bottom_card_num =
715 intptr_t(uintptr_t(_g1h->reserved_region().start()) >>
716 CardTableModRefBS::card_shift);
717
718 // Clear all the liveness counting data
719 clear_all_count_data();
720
721 // so that the call below can read a sensible value
722 _heap_start = (HeapWord*) heap_rs.base();
723 set_non_marking_state();
724 _completed_initialization = true;
725 }
726
727 void ConcurrentMark::update_g1_committed(bool force) {
728 // If concurrent marking is not in progress, then we do not need to
729 // update _heap_end.
730 if (!concurrent_marking_in_progress() && !force) return;
731
732 MemRegion committed = _g1h->g1_committed();
733 assert(committed.start() == _heap_start, "start shouldn't change");
734 HeapWord* new_end = committed.end();
735 if (new_end > _heap_end) {
736 // The heap has been expanded.
737
738 _heap_end = new_end;
739 }
740 // Notice that the heap can also shrink. However, this only happens
741 // during a Full GC (at least currently) and the entire marking
742 // phase will bail out and the task will not be restarted. So, let's
743 // do nothing.
744 }
745
746 void ConcurrentMark::reset() {
747 // Starting values for these two. This should be called in a STW
748 // phase. CM will be notified of any future g1_committed expansions
749 // will be at the end of evacuation pauses, when tasks are
750 // inactive.
751 MemRegion committed = _g1h->g1_committed();
752 _heap_start = committed.start();
753 _heap_end = committed.end();
754
755 // Separated the asserts so that we know which one fires.
756 assert(_heap_start != NULL, "heap bounds should look ok");
757 assert(_heap_end != NULL, "heap bounds should look ok");
758 assert(_heap_start < _heap_end, "heap bounds should look ok");
759
760 // Reset all the marking data structures and any necessary flags
761 reset_marking_state();
762
763 if (verbose_low()) {
764 gclog_or_tty->print_cr("[global] resetting");
765 }
766
767 // We do reset all of them, since different phases will use
768 // different number of active threads. So, it's easiest to have all
769 // of them ready.
770 for (uint i = 0; i < _max_worker_id; ++i) {
771 _tasks[i]->reset(_nextMarkBitMap);
772 }
773
774 // we need this to make sure that the flag is on during the evac
775 // pause with initial mark piggy-backed
776 set_concurrent_marking_in_progress();
777 }
778
779
780 void ConcurrentMark::reset_marking_state(bool clear_overflow) {
781 _markStack.set_should_expand();
782 _markStack.setEmpty(); // Also clears the _markStack overflow flag
783 if (clear_overflow) {
784 clear_has_overflown();
785 } else {
786 assert(has_overflown(), "pre-condition");
787 }
788 _finger = _heap_start;
789
790 for (uint i = 0; i < _max_worker_id; ++i) {
791 CMTaskQueue* queue = _task_queues->queue(i);
792 queue->set_empty();
793 }
794 }
795
796 void ConcurrentMark::set_concurrency(uint active_tasks) {
797 assert(active_tasks <= _max_worker_id, "we should not have more");
798
799 _active_tasks = active_tasks;
800 // Need to update the three data structures below according to the
801 // number of active threads for this phase.
802 _terminator = ParallelTaskTerminator((int) active_tasks, _task_queues);
803 _first_overflow_barrier_sync.set_n_workers((int) active_tasks);
804 _second_overflow_barrier_sync.set_n_workers((int) active_tasks);
805 }
806
807 void ConcurrentMark::set_concurrency_and_phase(uint active_tasks, bool concurrent) {
808 set_concurrency(active_tasks);
809
810 _concurrent = concurrent;
811 // We propagate this to all tasks, not just the active ones.
812 for (uint i = 0; i < _max_worker_id; ++i)
813 _tasks[i]->set_concurrent(concurrent);
814
815 if (concurrent) {
816 set_concurrent_marking_in_progress();
817 } else {
818 // We currently assume that the concurrent flag has been set to
819 // false before we start remark. At this point we should also be
820 // in a STW phase.
821 assert(!concurrent_marking_in_progress(), "invariant");
822 assert(out_of_regions(),
823 err_msg("only way to get here: _finger: "PTR_FORMAT", _heap_end: "PTR_FORMAT,
824 p2i(_finger), p2i(_heap_end)));
825 update_g1_committed(true);
826 }
827 }
828
829 void ConcurrentMark::set_non_marking_state() {
830 // We set the global marking state to some default values when we're
831 // not doing marking.
832 reset_marking_state();
833 _active_tasks = 0;
834 clear_concurrent_marking_in_progress();
835 }
836
837 ConcurrentMark::~ConcurrentMark() {
838 // The ConcurrentMark instance is never freed.
839 ShouldNotReachHere();
840 }
841
842 void ConcurrentMark::clearNextBitmap() {
843 G1CollectedHeap* g1h = G1CollectedHeap::heap();
844 G1CollectorPolicy* g1p = g1h->g1_policy();
845
846 // Make sure that the concurrent mark thread looks to still be in
847 // the current cycle.
848 guarantee(cmThread()->during_cycle(), "invariant");
849
850 // We are finishing up the current cycle by clearing the next
851 // marking bitmap and getting it ready for the next cycle. During
852 // this time no other cycle can start. So, let's make sure that this
853 // is the case.
854 guarantee(!g1h->mark_in_progress(), "invariant");
855
856 // clear the mark bitmap (no grey objects to start with).
857 // We need to do this in chunks and offer to yield in between
858 // each chunk.
859 HeapWord* start = _nextMarkBitMap->startWord();
860 HeapWord* end = _nextMarkBitMap->endWord();
861 HeapWord* cur = start;
862 size_t chunkSize = M;
863 while (cur < end) {
864 HeapWord* next = cur + chunkSize;
865 if (next > end) {
866 next = end;
867 }
868 MemRegion mr(cur,next);
869 _nextMarkBitMap->clearRange(mr);
870 cur = next;
871 do_yield_check();
872
873 // Repeat the asserts from above. We'll do them as asserts here to
874 // minimize their overhead on the product. However, we'll have
875 // them as guarantees at the beginning / end of the bitmap
876 // clearing to get some checking in the product.
877 assert(cmThread()->during_cycle(), "invariant");
878 assert(!g1h->mark_in_progress(), "invariant");
879 }
880
881 // Clear the liveness counting data
882 clear_all_count_data();
883
884 // Repeat the asserts from above.
885 guarantee(cmThread()->during_cycle(), "invariant");
886 guarantee(!g1h->mark_in_progress(), "invariant");
887 }
888
889 class NoteStartOfMarkHRClosure: public HeapRegionClosure {
890 public:
891 bool doHeapRegion(HeapRegion* r) {
892 if (!r->continuesHumongous()) {
893 r->note_start_of_marking();
894 }
895 return false;
896 }
897 };
898
899 void ConcurrentMark::checkpointRootsInitialPre() {
900 G1CollectedHeap* g1h = G1CollectedHeap::heap();
901 G1CollectorPolicy* g1p = g1h->g1_policy();
902
903 _has_aborted = false;
904
905 #ifndef PRODUCT
906 if (G1PrintReachableAtInitialMark) {
907 print_reachable("at-cycle-start",
908 VerifyOption_G1UsePrevMarking, true /* all */);
909 }
910 #endif
911
912 // Initialize marking structures. This has to be done in a STW phase.
913 reset();
914
915 // For each region note start of marking.
916 NoteStartOfMarkHRClosure startcl;
917 g1h->heap_region_iterate(&startcl);
918 }
919
920
921 void ConcurrentMark::checkpointRootsInitialPost() {
922 G1CollectedHeap* g1h = G1CollectedHeap::heap();
923
924 // If we force an overflow during remark, the remark operation will
925 // actually abort and we'll restart concurrent marking. If we always
926 // force an overflow during remark we'll never actually complete the
927 // marking phase. So, we initialize this here, at the start of the
928 // cycle, so that at the remaining overflow number will decrease at
929 // every remark and we'll eventually not need to cause one.
930 force_overflow_stw()->init();
931
932 // Start Concurrent Marking weak-reference discovery.
933 ReferenceProcessor* rp = g1h->ref_processor_cm();
934 // enable ("weak") refs discovery
935 rp->enable_discovery(true /*verify_disabled*/, true /*verify_no_refs*/);
936 rp->setup_policy(false); // snapshot the soft ref policy to be used in this cycle
937
938 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
939 // This is the start of the marking cycle, we're expected all
940 // threads to have SATB queues with active set to false.
941 satb_mq_set.set_active_all_threads(true, /* new active value */
942 false /* expected_active */);
943
944 _root_regions.prepare_for_scan();
945
946 // update_g1_committed() will be called at the end of an evac pause
947 // when marking is on. So, it's also called at the end of the
948 // initial-mark pause to update the heap end, if the heap expands
949 // during it. No need to call it here.
950 }
951
952 /*
953 * Notice that in the next two methods, we actually leave the STS
954 * during the barrier sync and join it immediately afterwards. If we
955 * do not do this, the following deadlock can occur: one thread could
956 * be in the barrier sync code, waiting for the other thread to also
957 * sync up, whereas another one could be trying to yield, while also
958 * waiting for the other threads to sync up too.
959 *
960 * Note, however, that this code is also used during remark and in
961 * this case we should not attempt to leave / enter the STS, otherwise
962 * we'll either hit an assert (debug / fastdebug) or deadlock
963 * (product). So we should only leave / enter the STS if we are
964 * operating concurrently.
965 *
966 * Because the thread that does the sync barrier has left the STS, it
967 * is possible to be suspended for a Full GC or an evacuation pause
968 * could occur. This is actually safe, since the entering the sync
969 * barrier is one of the last things do_marking_step() does, and it
970 * doesn't manipulate any data structures afterwards.
971 */
972
973 void ConcurrentMark::enter_first_sync_barrier(uint worker_id) {
974 if (verbose_low()) {
975 gclog_or_tty->print_cr("[%u] entering first barrier", worker_id);
976 }
977
978 if (concurrent()) {
979 SuspendibleThreadSet::leave();
980 }
981
982 bool barrier_aborted = !_first_overflow_barrier_sync.enter();
983
984 if (concurrent()) {
985 SuspendibleThreadSet::join();
986 }
987 // at this point everyone should have synced up and not be doing any
988 // more work
989
990 if (verbose_low()) {
991 if (barrier_aborted) {
992 gclog_or_tty->print_cr("[%u] aborted first barrier", worker_id);
993 } else {
994 gclog_or_tty->print_cr("[%u] leaving first barrier", worker_id);
995 }
996 }
997
998 if (barrier_aborted) {
999 // If the barrier aborted we ignore the overflow condition and
1000 // just abort the whole marking phase as quickly as possible.
1001 return;
1002 }
1003
1004 // If we're executing the concurrent phase of marking, reset the marking
1005 // state; otherwise the marking state is reset after reference processing,
1006 // during the remark pause.
1007 // If we reset here as a result of an overflow during the remark we will
1008 // see assertion failures from any subsequent set_concurrency_and_phase()
1009 // calls.
1010 if (concurrent()) {
1011 // let the task associated with with worker 0 do this
1012 if (worker_id == 0) {
1013 // task 0 is responsible for clearing the global data structures
1014 // We should be here because of an overflow. During STW we should
1015 // not clear the overflow flag since we rely on it being true when
1016 // we exit this method to abort the pause and restart concurrent
1017 // marking.
1018 reset_marking_state(true /* clear_overflow */);
1019 force_overflow()->update();
1020
1021 if (G1Log::fine()) {
1022 gclog_or_tty->date_stamp(PrintGCDateStamps);
1023 gclog_or_tty->stamp(PrintGCTimeStamps);
1024 gclog_or_tty->print_cr("[GC concurrent-mark-reset-for-overflow]");
1025 }
1026 }
1027 }
1028
1029 // after this, each task should reset its own data structures then
1030 // then go into the second barrier
1031 }
1032
1033 void ConcurrentMark::enter_second_sync_barrier(uint worker_id) {
1034 if (verbose_low()) {
1035 gclog_or_tty->print_cr("[%u] entering second barrier", worker_id);
1036 }
1037
1038 if (concurrent()) {
1039 SuspendibleThreadSet::leave();
1040 }
1041
1042 bool barrier_aborted = !_second_overflow_barrier_sync.enter();
1043
1044 if (concurrent()) {
1045 SuspendibleThreadSet::join();
1046 }
1047 // at this point everything should be re-initialized and ready to go
1048
1049 if (verbose_low()) {
1050 if (barrier_aborted) {
1051 gclog_or_tty->print_cr("[%u] aborted second barrier", worker_id);
1052 } else {
1053 gclog_or_tty->print_cr("[%u] leaving second barrier", worker_id);
1054 }
1055 }
1056 }
1057
1058 #ifndef PRODUCT
1059 void ForceOverflowSettings::init() {
1060 _num_remaining = G1ConcMarkForceOverflow;
1061 _force = false;
1062 update();
1063 }
1064
1065 void ForceOverflowSettings::update() {
1066 if (_num_remaining > 0) {
1067 _num_remaining -= 1;
1068 _force = true;
1069 } else {
1070 _force = false;
1071 }
1072 }
1073
1074 bool ForceOverflowSettings::should_force() {
1075 if (_force) {
1076 _force = false;
1077 return true;
1078 } else {
1079 return false;
1080 }
1081 }
1082 #endif // !PRODUCT
1083
1084 class CMConcurrentMarkingTask: public AbstractGangTask {
1085 private:
1086 ConcurrentMark* _cm;
1087 ConcurrentMarkThread* _cmt;
1088
1089 public:
1090 void work(uint worker_id) {
1091 assert(Thread::current()->is_ConcurrentGC_thread(),
1092 "this should only be done by a conc GC thread");
1093 ResourceMark rm;
1094
1095 double start_vtime = os::elapsedVTime();
1096
1097 SuspendibleThreadSet::join();
1098
1099 assert(worker_id < _cm->active_tasks(), "invariant");
1100 CMTask* the_task = _cm->task(worker_id);
1101 the_task->record_start_time();
1102 if (!_cm->has_aborted()) {
1103 do {
1104 double start_vtime_sec = os::elapsedVTime();
1105 double start_time_sec = os::elapsedTime();
1106 double mark_step_duration_ms = G1ConcMarkStepDurationMillis;
1107
1108 the_task->do_marking_step(mark_step_duration_ms,
1109 true /* do_termination */,
1110 false /* is_serial*/);
1111
1112 double end_time_sec = os::elapsedTime();
1113 double end_vtime_sec = os::elapsedVTime();
1114 double elapsed_vtime_sec = end_vtime_sec - start_vtime_sec;
1115 double elapsed_time_sec = end_time_sec - start_time_sec;
1116 _cm->clear_has_overflown();
1117
1118 bool ret = _cm->do_yield_check(worker_id);
1119
1120 jlong sleep_time_ms;
1121 if (!_cm->has_aborted() && the_task->has_aborted()) {
1122 sleep_time_ms =
1123 (jlong) (elapsed_vtime_sec * _cm->sleep_factor() * 1000.0);
1124 SuspendibleThreadSet::leave();
1125 os::sleep(Thread::current(), sleep_time_ms, false);
1126 SuspendibleThreadSet::join();
1127 }
1128 double end_time2_sec = os::elapsedTime();
1129 double elapsed_time2_sec = end_time2_sec - start_time_sec;
1130
1131 #if 0
1132 gclog_or_tty->print_cr("CM: elapsed %1.4lf ms, sleep %1.4lf ms, "
1133 "overhead %1.4lf",
1134 elapsed_vtime_sec * 1000.0, (double) sleep_time_ms,
1135 the_task->conc_overhead(os::elapsedTime()) * 8.0);
1136 gclog_or_tty->print_cr("elapsed time %1.4lf ms, time 2: %1.4lf ms",
1137 elapsed_time_sec * 1000.0, elapsed_time2_sec * 1000.0);
1138 #endif
1139 } while (!_cm->has_aborted() && the_task->has_aborted());
1140 }
1141 the_task->record_end_time();
1142 guarantee(!the_task->has_aborted() || _cm->has_aborted(), "invariant");
1143
1144 SuspendibleThreadSet::leave();
1145
1146 double end_vtime = os::elapsedVTime();
1147 _cm->update_accum_task_vtime(worker_id, end_vtime - start_vtime);
1148 }
1149
1150 CMConcurrentMarkingTask(ConcurrentMark* cm,
1151 ConcurrentMarkThread* cmt) :
1152 AbstractGangTask("Concurrent Mark"), _cm(cm), _cmt(cmt) { }
1153
1154 ~CMConcurrentMarkingTask() { }
1155 };
1156
1157 // Calculates the number of active workers for a concurrent
1158 // phase.
1159 uint ConcurrentMark::calc_parallel_marking_threads() {
1160 if (G1CollectedHeap::use_parallel_gc_threads()) {
1161 uint n_conc_workers = 0;
1162 if (!UseDynamicNumberOfGCThreads ||
1163 (!FLAG_IS_DEFAULT(ConcGCThreads) &&
1164 !ForceDynamicNumberOfGCThreads)) {
1165 n_conc_workers = max_parallel_marking_threads();
1166 } else {
1167 n_conc_workers =
1168 AdaptiveSizePolicy::calc_default_active_workers(
1169 max_parallel_marking_threads(),
1170 1, /* Minimum workers */
1171 parallel_marking_threads(),
1172 Threads::number_of_non_daemon_threads());
1173 // Don't scale down "n_conc_workers" by scale_parallel_threads() because
1174 // that scaling has already gone into "_max_parallel_marking_threads".
1175 }
1176 assert(n_conc_workers > 0, "Always need at least 1");
1177 return n_conc_workers;
1178 }
1179 // If we are not running with any parallel GC threads we will not
1180 // have spawned any marking threads either. Hence the number of
1181 // concurrent workers should be 0.
1182 return 0;
1183 }
1184
1185 void ConcurrentMark::scanRootRegion(HeapRegion* hr, uint worker_id) {
1186 // Currently, only survivors can be root regions.
1187 assert(hr->next_top_at_mark_start() == hr->bottom(), "invariant");
1188 G1RootRegionScanClosure cl(_g1h, this, worker_id);
1189
1190 const uintx interval = PrefetchScanIntervalInBytes;
1191 HeapWord* curr = hr->bottom();
1192 const HeapWord* end = hr->top();
1193 while (curr < end) {
1194 Prefetch::read(curr, interval);
1195 oop obj = oop(curr);
1196 int size = obj->oop_iterate(&cl);
1197 assert(size == obj->size(), "sanity");
1198 curr += size;
1199 }
1200 }
1201
1202 class CMRootRegionScanTask : public AbstractGangTask {
1203 private:
1204 ConcurrentMark* _cm;
1205
1206 public:
1207 CMRootRegionScanTask(ConcurrentMark* cm) :
1208 AbstractGangTask("Root Region Scan"), _cm(cm) { }
1209
1210 void work(uint worker_id) {
1211 assert(Thread::current()->is_ConcurrentGC_thread(),
1212 "this should only be done by a conc GC thread");
1213
1214 CMRootRegions* root_regions = _cm->root_regions();
1215 HeapRegion* hr = root_regions->claim_next();
1216 while (hr != NULL) {
1217 _cm->scanRootRegion(hr, worker_id);
1218 hr = root_regions->claim_next();
1219 }
1220 }
1221 };
1222
1223 void ConcurrentMark::scanRootRegions() {
1224 // scan_in_progress() will have been set to true only if there was
1225 // at least one root region to scan. So, if it's false, we
1226 // should not attempt to do any further work.
1227 if (root_regions()->scan_in_progress()) {
1228 _parallel_marking_threads = calc_parallel_marking_threads();
1229 assert(parallel_marking_threads() <= max_parallel_marking_threads(),
1230 "Maximum number of marking threads exceeded");
1231 uint active_workers = MAX2(1U, parallel_marking_threads());
1232
1233 CMRootRegionScanTask task(this);
1234 if (use_parallel_marking_threads()) {
1235 _parallel_workers->set_active_workers((int) active_workers);
1236 _parallel_workers->run_task(&task);
1237 } else {
1238 task.work(0);
1239 }
1240
1241 // It's possible that has_aborted() is true here without actually
1242 // aborting the survivor scan earlier. This is OK as it's
1243 // mainly used for sanity checking.
1244 root_regions()->scan_finished();
1245 }
1246 }
1247
1248 void ConcurrentMark::markFromRoots() {
1249 // we might be tempted to assert that:
1250 // assert(asynch == !SafepointSynchronize::is_at_safepoint(),
1251 // "inconsistent argument?");
1252 // However that wouldn't be right, because it's possible that
1253 // a safepoint is indeed in progress as a younger generation
1254 // stop-the-world GC happens even as we mark in this generation.
1255
1256 _restart_for_overflow = false;
1257 force_overflow_conc()->init();
1258
1259 // _g1h has _n_par_threads
1260 _parallel_marking_threads = calc_parallel_marking_threads();
1261 assert(parallel_marking_threads() <= max_parallel_marking_threads(),
1262 "Maximum number of marking threads exceeded");
1263
1264 uint active_workers = MAX2(1U, parallel_marking_threads());
1265
1266 // Parallel task terminator is set in "set_concurrency_and_phase()"
1267 set_concurrency_and_phase(active_workers, true /* concurrent */);
1268
1269 CMConcurrentMarkingTask markingTask(this, cmThread());
1270 if (use_parallel_marking_threads()) {
1271 _parallel_workers->set_active_workers((int)active_workers);
1272 // Don't set _n_par_threads because it affects MT in process_strong_roots()
1273 // and the decisions on that MT processing is made elsewhere.
1274 assert(_parallel_workers->active_workers() > 0, "Should have been set");
1275 _parallel_workers->run_task(&markingTask);
1276 } else {
1277 markingTask.work(0);
1278 }
1279 print_stats();
1280 }
1281
1282 void ConcurrentMark::checkpointRootsFinal(bool clear_all_soft_refs) {
1283 // world is stopped at this checkpoint
1284 assert(SafepointSynchronize::is_at_safepoint(),
1285 "world should be stopped");
1286
1287 G1CollectedHeap* g1h = G1CollectedHeap::heap();
1288
1289 // If a full collection has happened, we shouldn't do this.
1290 if (has_aborted()) {
1291 g1h->set_marking_complete(); // So bitmap clearing isn't confused
1292 return;
1293 }
1294
1295 SvcGCMarker sgcm(SvcGCMarker::OTHER);
1296
1297 if (VerifyDuringGC) {
1298 HandleMark hm; // handle scope
1299 Universe::heap()->prepare_for_verify();
1300 Universe::verify(VerifyOption_G1UsePrevMarking,
1301 " VerifyDuringGC:(before)");
1302 }
1303 g1h->check_bitmaps("Remark Start");
1304
1305 G1CollectorPolicy* g1p = g1h->g1_policy();
1306 g1p->record_concurrent_mark_remark_start();
1307
1308 double start = os::elapsedTime();
1309
1310 checkpointRootsFinalWork();
1311
1312 double mark_work_end = os::elapsedTime();
1313
1314 weakRefsWork(clear_all_soft_refs);
1315
1316 if (has_overflown()) {
1317 // Oops. We overflowed. Restart concurrent marking.
1318 _restart_for_overflow = true;
1319 if (G1TraceMarkStackOverflow) {
1320 gclog_or_tty->print_cr("\nRemark led to restart for overflow.");
1321 }
1322
1323 // Verify the heap w.r.t. the previous marking bitmap.
1324 if (VerifyDuringGC) {
1325 HandleMark hm; // handle scope
1326 Universe::heap()->prepare_for_verify();
1327 Universe::verify(VerifyOption_G1UsePrevMarking,
1328 " VerifyDuringGC:(overflow)");
1329 }
1330
1331 // Clear the marking state because we will be restarting
1332 // marking due to overflowing the global mark stack.
1333 reset_marking_state();
1334 } else {
1335 // Aggregate the per-task counting data that we have accumulated
1336 // while marking.
1337 aggregate_count_data();
1338
1339 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
1340 // We're done with marking.
1341 // This is the end of the marking cycle, we're expected all
1342 // threads to have SATB queues with active set to true.
1343 satb_mq_set.set_active_all_threads(false, /* new active value */
1344 true /* expected_active */);
1345
1346 if (VerifyDuringGC) {
1347 HandleMark hm; // handle scope
1348 Universe::heap()->prepare_for_verify();
1349 Universe::verify(VerifyOption_G1UseNextMarking,
1350 " VerifyDuringGC:(after)");
1351 }
1352 g1h->check_bitmaps("Remark End");
1353 assert(!restart_for_overflow(), "sanity");
1354 // Completely reset the marking state since marking completed
1355 set_non_marking_state();
1356 }
1357
1358 // Expand the marking stack, if we have to and if we can.
1359 if (_markStack.should_expand()) {
1360 _markStack.expand();
1361 }
1362
1363 // Statistics
1364 double now = os::elapsedTime();
1365 _remark_mark_times.add((mark_work_end - start) * 1000.0);
1366 _remark_weak_ref_times.add((now - mark_work_end) * 1000.0);
1367 _remark_times.add((now - start) * 1000.0);
1368
1369 g1p->record_concurrent_mark_remark_end();
1370
1371 G1CMIsAliveClosure is_alive(g1h);
1372 g1h->gc_tracer_cm()->report_object_count_after_gc(&is_alive);
1373 }
1374
1375 // Base class of the closures that finalize and verify the
1376 // liveness counting data.
1377 class CMCountDataClosureBase: public HeapRegionClosure {
1378 protected:
1379 G1CollectedHeap* _g1h;
1380 ConcurrentMark* _cm;
1381 CardTableModRefBS* _ct_bs;
1382
1383 BitMap* _region_bm;
1384 BitMap* _card_bm;
1385
1386 // Takes a region that's not empty (i.e., it has at least one
1387 // live object in it and sets its corresponding bit on the region
1388 // bitmap to 1. If the region is "starts humongous" it will also set
1389 // to 1 the bits on the region bitmap that correspond to its
1390 // associated "continues humongous" regions.
1391 void set_bit_for_region(HeapRegion* hr) {
1392 assert(!hr->continuesHumongous(), "should have filtered those out");
1393
1394 BitMap::idx_t index = (BitMap::idx_t) hr->hrs_index();
1395 if (!hr->startsHumongous()) {
1396 // Normal (non-humongous) case: just set the bit.
1397 _region_bm->par_at_put(index, true);
1398 } else {
1399 // Starts humongous case: calculate how many regions are part of
1400 // this humongous region and then set the bit range.
1401 BitMap::idx_t end_index = (BitMap::idx_t) hr->last_hc_index();
1402 _region_bm->par_at_put_range(index, end_index, true);
1403 }
1404 }
1405
1406 public:
1407 CMCountDataClosureBase(G1CollectedHeap* g1h,
1408 BitMap* region_bm, BitMap* card_bm):
1409 _g1h(g1h), _cm(g1h->concurrent_mark()),
1410 _ct_bs((CardTableModRefBS*) (g1h->barrier_set())),
1411 _region_bm(region_bm), _card_bm(card_bm) { }
1412 };
1413
1414 // Closure that calculates the # live objects per region. Used
1415 // for verification purposes during the cleanup pause.
1416 class CalcLiveObjectsClosure: public CMCountDataClosureBase {
1417 CMBitMapRO* _bm;
1418 size_t _region_marked_bytes;
1419
1420 public:
1421 CalcLiveObjectsClosure(CMBitMapRO *bm, G1CollectedHeap* g1h,
1422 BitMap* region_bm, BitMap* card_bm) :
1423 CMCountDataClosureBase(g1h, region_bm, card_bm),
1424 _bm(bm), _region_marked_bytes(0) { }
1425
1426 bool doHeapRegion(HeapRegion* hr) {
1427
1428 if (hr->continuesHumongous()) {
1429 // We will ignore these here and process them when their
1430 // associated "starts humongous" region is processed (see
1431 // set_bit_for_heap_region()). Note that we cannot rely on their
1432 // associated "starts humongous" region to have their bit set to
1433 // 1 since, due to the region chunking in the parallel region
1434 // iteration, a "continues humongous" region might be visited
1435 // before its associated "starts humongous".
1436 return false;
1437 }
1438
1439 HeapWord* ntams = hr->next_top_at_mark_start();
1440 HeapWord* start = hr->bottom();
1441
1442 assert(start <= hr->end() && start <= ntams && ntams <= hr->end(),
1443 err_msg("Preconditions not met - "
1444 "start: "PTR_FORMAT", ntams: "PTR_FORMAT", end: "PTR_FORMAT,
1445 p2i(start), p2i(ntams), p2i(hr->end())));
1446
1447 // Find the first marked object at or after "start".
1448 start = _bm->getNextMarkedWordAddress(start, ntams);
1449
1450 size_t marked_bytes = 0;
1451
1452 while (start < ntams) {
1453 oop obj = oop(start);
1454 int obj_sz = obj->size();
1455 HeapWord* obj_end = start + obj_sz;
1456
1457 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(start);
1458 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(obj_end);
1459
1460 // Note: if we're looking at the last region in heap - obj_end
1461 // could be actually just beyond the end of the heap; end_idx
1462 // will then correspond to a (non-existent) card that is also
1463 // just beyond the heap.
1464 if (_g1h->is_in_g1_reserved(obj_end) && !_ct_bs->is_card_aligned(obj_end)) {
1465 // end of object is not card aligned - increment to cover
1466 // all the cards spanned by the object
1467 end_idx += 1;
1468 }
1469
1470 // Set the bits in the card BM for the cards spanned by this object.
1471 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1472
1473 // Add the size of this object to the number of marked bytes.
1474 marked_bytes += (size_t)obj_sz * HeapWordSize;
1475
1476 // Find the next marked object after this one.
1477 start = _bm->getNextMarkedWordAddress(obj_end, ntams);
1478 }
1479
1480 // Mark the allocated-since-marking portion...
1481 HeapWord* top = hr->top();
1482 if (ntams < top) {
1483 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(ntams);
1484 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(top);
1485
1486 // Note: if we're looking at the last region in heap - top
1487 // could be actually just beyond the end of the heap; end_idx
1488 // will then correspond to a (non-existent) card that is also
1489 // just beyond the heap.
1490 if (_g1h->is_in_g1_reserved(top) && !_ct_bs->is_card_aligned(top)) {
1491 // end of object is not card aligned - increment to cover
1492 // all the cards spanned by the object
1493 end_idx += 1;
1494 }
1495 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1496
1497 // This definitely means the region has live objects.
1498 set_bit_for_region(hr);
1499 }
1500
1501 // Update the live region bitmap.
1502 if (marked_bytes > 0) {
1503 set_bit_for_region(hr);
1504 }
1505
1506 // Set the marked bytes for the current region so that
1507 // it can be queried by a calling verification routine
1508 _region_marked_bytes = marked_bytes;
1509
1510 return false;
1511 }
1512
1513 size_t region_marked_bytes() const { return _region_marked_bytes; }
1514 };
1515
1516 // Heap region closure used for verifying the counting data
1517 // that was accumulated concurrently and aggregated during
1518 // the remark pause. This closure is applied to the heap
1519 // regions during the STW cleanup pause.
1520
1521 class VerifyLiveObjectDataHRClosure: public HeapRegionClosure {
1522 G1CollectedHeap* _g1h;
1523 ConcurrentMark* _cm;
1524 CalcLiveObjectsClosure _calc_cl;
1525 BitMap* _region_bm; // Region BM to be verified
1526 BitMap* _card_bm; // Card BM to be verified
1527 bool _verbose; // verbose output?
1528
1529 BitMap* _exp_region_bm; // Expected Region BM values
1530 BitMap* _exp_card_bm; // Expected card BM values
1531
1532 int _failures;
1533
1534 public:
1535 VerifyLiveObjectDataHRClosure(G1CollectedHeap* g1h,
1536 BitMap* region_bm,
1537 BitMap* card_bm,
1538 BitMap* exp_region_bm,
1539 BitMap* exp_card_bm,
1540 bool verbose) :
1541 _g1h(g1h), _cm(g1h->concurrent_mark()),
1542 _calc_cl(_cm->nextMarkBitMap(), g1h, exp_region_bm, exp_card_bm),
1543 _region_bm(region_bm), _card_bm(card_bm), _verbose(verbose),
1544 _exp_region_bm(exp_region_bm), _exp_card_bm(exp_card_bm),
1545 _failures(0) { }
1546
1547 int failures() const { return _failures; }
1548
1549 bool doHeapRegion(HeapRegion* hr) {
1550 if (hr->continuesHumongous()) {
1551 // We will ignore these here and process them when their
1552 // associated "starts humongous" region is processed (see
1553 // set_bit_for_heap_region()). Note that we cannot rely on their
1554 // associated "starts humongous" region to have their bit set to
1555 // 1 since, due to the region chunking in the parallel region
1556 // iteration, a "continues humongous" region might be visited
1557 // before its associated "starts humongous".
1558 return false;
1559 }
1560
1561 int failures = 0;
1562
1563 // Call the CalcLiveObjectsClosure to walk the marking bitmap for
1564 // this region and set the corresponding bits in the expected region
1565 // and card bitmaps.
1566 bool res = _calc_cl.doHeapRegion(hr);
1567 assert(res == false, "should be continuing");
1568
1569 MutexLockerEx x((_verbose ? ParGCRareEvent_lock : NULL),
1570 Mutex::_no_safepoint_check_flag);
1571
1572 // Verify the marked bytes for this region.
1573 size_t exp_marked_bytes = _calc_cl.region_marked_bytes();
1574 size_t act_marked_bytes = hr->next_marked_bytes();
1575
1576 // We're not OK if expected marked bytes > actual marked bytes. It means
1577 // we have missed accounting some objects during the actual marking.
1578 if (exp_marked_bytes > act_marked_bytes) {
1579 if (_verbose) {
1580 gclog_or_tty->print_cr("Region %u: marked bytes mismatch: "
1581 "expected: " SIZE_FORMAT ", actual: " SIZE_FORMAT,
1582 hr->hrs_index(), exp_marked_bytes, act_marked_bytes);
1583 }
1584 failures += 1;
1585 }
1586
1587 // Verify the bit, for this region, in the actual and expected
1588 // (which was just calculated) region bit maps.
1589 // We're not OK if the bit in the calculated expected region
1590 // bitmap is set and the bit in the actual region bitmap is not.
1591 BitMap::idx_t index = (BitMap::idx_t) hr->hrs_index();
1592
1593 bool expected = _exp_region_bm->at(index);
1594 bool actual = _region_bm->at(index);
1595 if (expected && !actual) {
1596 if (_verbose) {
1597 gclog_or_tty->print_cr("Region %u: region bitmap mismatch: "
1598 "expected: %s, actual: %s",
1599 hr->hrs_index(),
1600 BOOL_TO_STR(expected), BOOL_TO_STR(actual));
1601 }
1602 failures += 1;
1603 }
1604
1605 // Verify that the card bit maps for the cards spanned by the current
1606 // region match. We have an error if we have a set bit in the expected
1607 // bit map and the corresponding bit in the actual bitmap is not set.
1608
1609 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(hr->bottom());
1610 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(hr->top());
1611
1612 for (BitMap::idx_t i = start_idx; i < end_idx; i+=1) {
1613 expected = _exp_card_bm->at(i);
1614 actual = _card_bm->at(i);
1615
1616 if (expected && !actual) {
1617 if (_verbose) {
1618 gclog_or_tty->print_cr("Region %u: card bitmap mismatch at " SIZE_FORMAT ": "
1619 "expected: %s, actual: %s",
1620 hr->hrs_index(), i,
1621 BOOL_TO_STR(expected), BOOL_TO_STR(actual));
1622 }
1623 failures += 1;
1624 }
1625 }
1626
1627 if (failures > 0 && _verbose) {
1628 gclog_or_tty->print_cr("Region " HR_FORMAT ", ntams: " PTR_FORMAT ", "
1629 "marked_bytes: calc/actual " SIZE_FORMAT "/" SIZE_FORMAT,
1630 HR_FORMAT_PARAMS(hr), p2i(hr->next_top_at_mark_start()),
1631 _calc_cl.region_marked_bytes(), hr->next_marked_bytes());
1632 }
1633
1634 _failures += failures;
1635
1636 // We could stop iteration over the heap when we
1637 // find the first violating region by returning true.
1638 return false;
1639 }
1640 };
1641
1642 class G1ParVerifyFinalCountTask: public AbstractGangTask {
1643 protected:
1644 G1CollectedHeap* _g1h;
1645 ConcurrentMark* _cm;
1646 BitMap* _actual_region_bm;
1647 BitMap* _actual_card_bm;
1648
1649 uint _n_workers;
1650
1651 BitMap* _expected_region_bm;
1652 BitMap* _expected_card_bm;
1653
1654 int _failures;
1655 bool _verbose;
1656
1657 public:
1658 G1ParVerifyFinalCountTask(G1CollectedHeap* g1h,
1659 BitMap* region_bm, BitMap* card_bm,
1660 BitMap* expected_region_bm, BitMap* expected_card_bm)
1661 : AbstractGangTask("G1 verify final counting"),
1662 _g1h(g1h), _cm(_g1h->concurrent_mark()),
1663 _actual_region_bm(region_bm), _actual_card_bm(card_bm),
1664 _expected_region_bm(expected_region_bm), _expected_card_bm(expected_card_bm),
1665 _failures(0), _verbose(false),
1666 _n_workers(0) {
1667 assert(VerifyDuringGC, "don't call this otherwise");
1668
1669 // Use the value already set as the number of active threads
1670 // in the call to run_task().
1671 if (G1CollectedHeap::use_parallel_gc_threads()) {
1672 assert( _g1h->workers()->active_workers() > 0,
1673 "Should have been previously set");
1674 _n_workers = _g1h->workers()->active_workers();
1675 } else {
1676 _n_workers = 1;
1677 }
1678
1679 assert(_expected_card_bm->size() == _actual_card_bm->size(), "sanity");
1680 assert(_expected_region_bm->size() == _actual_region_bm->size(), "sanity");
1681
1682 _verbose = _cm->verbose_medium();
1683 }
1684
1685 void work(uint worker_id) {
1686 assert(worker_id < _n_workers, "invariant");
1687
1688 VerifyLiveObjectDataHRClosure verify_cl(_g1h,
1689 _actual_region_bm, _actual_card_bm,
1690 _expected_region_bm,
1691 _expected_card_bm,
1692 _verbose);
1693
1694 if (G1CollectedHeap::use_parallel_gc_threads()) {
1695 _g1h->heap_region_par_iterate_chunked(&verify_cl,
1696 worker_id,
1697 _n_workers,
1698 HeapRegion::VerifyCountClaimValue);
1699 } else {
1700 _g1h->heap_region_iterate(&verify_cl);
1701 }
1702
1703 Atomic::add(verify_cl.failures(), &_failures);
1704 }
1705
1706 int failures() const { return _failures; }
1707 };
1708
1709 // Closure that finalizes the liveness counting data.
1710 // Used during the cleanup pause.
1711 // Sets the bits corresponding to the interval [NTAMS, top]
1712 // (which contains the implicitly live objects) in the
1713 // card liveness bitmap. Also sets the bit for each region,
1714 // containing live data, in the region liveness bitmap.
1715
1716 class FinalCountDataUpdateClosure: public CMCountDataClosureBase {
1717 public:
1718 FinalCountDataUpdateClosure(G1CollectedHeap* g1h,
1719 BitMap* region_bm,
1720 BitMap* card_bm) :
1721 CMCountDataClosureBase(g1h, region_bm, card_bm) { }
1722
1723 bool doHeapRegion(HeapRegion* hr) {
1724
1725 if (hr->continuesHumongous()) {
1726 // We will ignore these here and process them when their
1727 // associated "starts humongous" region is processed (see
1728 // set_bit_for_heap_region()). Note that we cannot rely on their
1729 // associated "starts humongous" region to have their bit set to
1730 // 1 since, due to the region chunking in the parallel region
1731 // iteration, a "continues humongous" region might be visited
1732 // before its associated "starts humongous".
1733 return false;
1734 }
1735
1736 HeapWord* ntams = hr->next_top_at_mark_start();
1737 HeapWord* top = hr->top();
1738
1739 assert(hr->bottom() <= ntams && ntams <= hr->end(), "Preconditions.");
1740
1741 // Mark the allocated-since-marking portion...
1742 if (ntams < top) {
1743 // This definitely means the region has live objects.
1744 set_bit_for_region(hr);
1745
1746 // Now set the bits in the card bitmap for [ntams, top)
1747 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(ntams);
1748 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(top);
1749
1750 // Note: if we're looking at the last region in heap - top
1751 // could be actually just beyond the end of the heap; end_idx
1752 // will then correspond to a (non-existent) card that is also
1753 // just beyond the heap.
1754 if (_g1h->is_in_g1_reserved(top) && !_ct_bs->is_card_aligned(top)) {
1755 // end of object is not card aligned - increment to cover
1756 // all the cards spanned by the object
1757 end_idx += 1;
1758 }
1759
1760 assert(end_idx <= _card_bm->size(),
1761 err_msg("oob: end_idx= "SIZE_FORMAT", bitmap size= "SIZE_FORMAT,
1762 end_idx, _card_bm->size()));
1763 assert(start_idx < _card_bm->size(),
1764 err_msg("oob: start_idx= "SIZE_FORMAT", bitmap size= "SIZE_FORMAT,
1765 start_idx, _card_bm->size()));
1766
1767 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1768 }
1769
1770 // Set the bit for the region if it contains live data
1771 if (hr->next_marked_bytes() > 0) {
1772 set_bit_for_region(hr);
1773 }
1774
1775 return false;
1776 }
1777 };
1778
1779 class G1ParFinalCountTask: public AbstractGangTask {
1780 protected:
1781 G1CollectedHeap* _g1h;
1782 ConcurrentMark* _cm;
1783 BitMap* _actual_region_bm;
1784 BitMap* _actual_card_bm;
1785
1786 uint _n_workers;
1787
1788 public:
1789 G1ParFinalCountTask(G1CollectedHeap* g1h, BitMap* region_bm, BitMap* card_bm)
1790 : AbstractGangTask("G1 final counting"),
1791 _g1h(g1h), _cm(_g1h->concurrent_mark()),
1792 _actual_region_bm(region_bm), _actual_card_bm(card_bm),
1793 _n_workers(0) {
1794 // Use the value already set as the number of active threads
1795 // in the call to run_task().
1796 if (G1CollectedHeap::use_parallel_gc_threads()) {
1797 assert( _g1h->workers()->active_workers() > 0,
1798 "Should have been previously set");
1799 _n_workers = _g1h->workers()->active_workers();
1800 } else {
1801 _n_workers = 1;
1802 }
1803 }
1804
1805 void work(uint worker_id) {
1806 assert(worker_id < _n_workers, "invariant");
1807
1808 FinalCountDataUpdateClosure final_update_cl(_g1h,
1809 _actual_region_bm,
1810 _actual_card_bm);
1811
1812 if (G1CollectedHeap::use_parallel_gc_threads()) {
1813 _g1h->heap_region_par_iterate_chunked(&final_update_cl,
1814 worker_id,
1815 _n_workers,
1816 HeapRegion::FinalCountClaimValue);
1817 } else {
1818 _g1h->heap_region_iterate(&final_update_cl);
1819 }
1820 }
1821 };
1822
1823 class G1ParNoteEndTask;
1824
1825 class G1NoteEndOfConcMarkClosure : public HeapRegionClosure {
1826 G1CollectedHeap* _g1;
1827 size_t _max_live_bytes;
1828 uint _regions_claimed;
1829 size_t _freed_bytes;
1830 FreeRegionList* _local_cleanup_list;
1831 HeapRegionSetCount _old_regions_removed;
1832 HeapRegionSetCount _humongous_regions_removed;
1833 HRRSCleanupTask* _hrrs_cleanup_task;
1834 double _claimed_region_time;
1835 double _max_region_time;
1836
1837 public:
1838 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1,
1839 FreeRegionList* local_cleanup_list,
1840 HRRSCleanupTask* hrrs_cleanup_task) :
1841 _g1(g1),
1842 _max_live_bytes(0), _regions_claimed(0),
1843 _freed_bytes(0),
1844 _claimed_region_time(0.0), _max_region_time(0.0),
1845 _local_cleanup_list(local_cleanup_list),
1846 _old_regions_removed(),
1847 _humongous_regions_removed(),
1848 _hrrs_cleanup_task(hrrs_cleanup_task) { }
1849
1850 size_t freed_bytes() { return _freed_bytes; }
1851 const HeapRegionSetCount& old_regions_removed() { return _old_regions_removed; }
1852 const HeapRegionSetCount& humongous_regions_removed() { return _humongous_regions_removed; }
1853
1854 bool doHeapRegion(HeapRegion *hr) {
1855 if (hr->continuesHumongous()) {
1856 return false;
1857 }
1858 // We use a claim value of zero here because all regions
1859 // were claimed with value 1 in the FinalCount task.
1860 _g1->reset_gc_time_stamps(hr);
1861 double start = os::elapsedTime();
1862 _regions_claimed++;
1863 hr->note_end_of_marking();
1864 _max_live_bytes += hr->max_live_bytes();
1865
1866 if (hr->used() > 0 && hr->max_live_bytes() == 0 && !hr->is_young()) {
1867 _freed_bytes += hr->used();
1868 hr->set_containing_set(NULL);
1869 if (hr->isHumongous()) {
1870 assert(hr->startsHumongous(), "we should only see starts humongous");
1871 _humongous_regions_removed.increment(1u, hr->capacity());
1872 _g1->free_humongous_region(hr, _local_cleanup_list, true);
1873 } else {
1874 _old_regions_removed.increment(1u, hr->capacity());
1875 _g1->free_region(hr, _local_cleanup_list, true);
1876 }
1877 } else {
1878 hr->rem_set()->do_cleanup_work(_hrrs_cleanup_task);
1879 }
1880
1881 double region_time = (os::elapsedTime() - start);
1882 _claimed_region_time += region_time;
1883 if (region_time > _max_region_time) {
1884 _max_region_time = region_time;
1885 }
1886 return false;
1887 }
1888
1889 size_t max_live_bytes() { return _max_live_bytes; }
1890 uint regions_claimed() { return _regions_claimed; }
1891 double claimed_region_time_sec() { return _claimed_region_time; }
1892 double max_region_time_sec() { return _max_region_time; }
1893 };
1894
1895 class G1ParNoteEndTask: public AbstractGangTask {
1896 friend class G1NoteEndOfConcMarkClosure;
1897
1898 protected:
1899 G1CollectedHeap* _g1h;
1900 size_t _max_live_bytes;
1901 size_t _freed_bytes;
1902 FreeRegionList* _cleanup_list;
1903
1904 public:
1905 G1ParNoteEndTask(G1CollectedHeap* g1h,
1906 FreeRegionList* cleanup_list) :
1907 AbstractGangTask("G1 note end"), _g1h(g1h),
1908 _max_live_bytes(0), _freed_bytes(0), _cleanup_list(cleanup_list) { }
1909
1910 void work(uint worker_id) {
1911 double start = os::elapsedTime();
1912 FreeRegionList local_cleanup_list("Local Cleanup List");
1913 HRRSCleanupTask hrrs_cleanup_task;
1914 G1NoteEndOfConcMarkClosure g1_note_end(_g1h, &local_cleanup_list,
1915 &hrrs_cleanup_task);
1916 if (G1CollectedHeap::use_parallel_gc_threads()) {
1917 _g1h->heap_region_par_iterate_chunked(&g1_note_end, worker_id,
1918 _g1h->workers()->active_workers(),
1919 HeapRegion::NoteEndClaimValue);
1920 } else {
1921 _g1h->heap_region_iterate(&g1_note_end);
1922 }
1923 assert(g1_note_end.complete(), "Shouldn't have yielded!");
1924
1925 // Now update the lists
1926 _g1h->remove_from_old_sets(g1_note_end.old_regions_removed(), g1_note_end.humongous_regions_removed());
1927 {
1928 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
1929 _g1h->decrement_summary_bytes(g1_note_end.freed_bytes());
1930 _max_live_bytes += g1_note_end.max_live_bytes();
1931 _freed_bytes += g1_note_end.freed_bytes();
1932
1933 // If we iterate over the global cleanup list at the end of
1934 // cleanup to do this printing we will not guarantee to only
1935 // generate output for the newly-reclaimed regions (the list
1936 // might not be empty at the beginning of cleanup; we might
1937 // still be working on its previous contents). So we do the
1938 // printing here, before we append the new regions to the global
1939 // cleanup list.
1940
1941 G1HRPrinter* hr_printer = _g1h->hr_printer();
1942 if (hr_printer->is_active()) {
1943 FreeRegionListIterator iter(&local_cleanup_list);
1944 while (iter.more_available()) {
1945 HeapRegion* hr = iter.get_next();
1946 hr_printer->cleanup(hr);
1947 }
1948 }
1949
1950 _cleanup_list->add_ordered(&local_cleanup_list);
1951 assert(local_cleanup_list.is_empty(), "post-condition");
1952
1953 HeapRegionRemSet::finish_cleanup_task(&hrrs_cleanup_task);
1954 }
1955 }
1956 size_t max_live_bytes() { return _max_live_bytes; }
1957 size_t freed_bytes() { return _freed_bytes; }
1958 };
1959
1960 class G1ParScrubRemSetTask: public AbstractGangTask {
1961 protected:
1962 G1RemSet* _g1rs;
1963 BitMap* _region_bm;
1964 BitMap* _card_bm;
1965 public:
1966 G1ParScrubRemSetTask(G1CollectedHeap* g1h,
1967 BitMap* region_bm, BitMap* card_bm) :
1968 AbstractGangTask("G1 ScrubRS"), _g1rs(g1h->g1_rem_set()),
1969 _region_bm(region_bm), _card_bm(card_bm) { }
1970
1971 void work(uint worker_id) {
1972 if (G1CollectedHeap::use_parallel_gc_threads()) {
1973 _g1rs->scrub_par(_region_bm, _card_bm, worker_id,
1974 HeapRegion::ScrubRemSetClaimValue);
1975 } else {
1976 _g1rs->scrub(_region_bm, _card_bm);
1977 }
1978 }
1979
1980 };
1981
1982 void ConcurrentMark::cleanup() {
1983 // world is stopped at this checkpoint
1984 assert(SafepointSynchronize::is_at_safepoint(),
1985 "world should be stopped");
1986 G1CollectedHeap* g1h = G1CollectedHeap::heap();
1987
1988 // If a full collection has happened, we shouldn't do this.
1989 if (has_aborted()) {
1990 g1h->set_marking_complete(); // So bitmap clearing isn't confused
1991 return;
1992 }
1993
1994 g1h->verify_region_sets_optional();
1995
1996 if (VerifyDuringGC) {
1997 HandleMark hm; // handle scope
1998 Universe::heap()->prepare_for_verify();
1999 Universe::verify(VerifyOption_G1UsePrevMarking,
2000 " VerifyDuringGC:(before)");
2001 }
2002 g1h->check_bitmaps("Cleanup Start");
2003
2004 G1CollectorPolicy* g1p = G1CollectedHeap::heap()->g1_policy();
2005 g1p->record_concurrent_mark_cleanup_start();
2006
2007 double start = os::elapsedTime();
2008
2009 HeapRegionRemSet::reset_for_cleanup_tasks();
2010
2011 uint n_workers;
2012
2013 // Do counting once more with the world stopped for good measure.
2014 G1ParFinalCountTask g1_par_count_task(g1h, &_region_bm, &_card_bm);
2015
2016 if (G1CollectedHeap::use_parallel_gc_threads()) {
2017 assert(g1h->check_heap_region_claim_values(HeapRegion::InitialClaimValue),
2018 "sanity check");
2019
2020 g1h->set_par_threads();
2021 n_workers = g1h->n_par_threads();
2022 assert(g1h->n_par_threads() == n_workers,
2023 "Should not have been reset");
2024 g1h->workers()->run_task(&g1_par_count_task);
2025 // Done with the parallel phase so reset to 0.
2026 g1h->set_par_threads(0);
2027
2028 assert(g1h->check_heap_region_claim_values(HeapRegion::FinalCountClaimValue),
2029 "sanity check");
2030 } else {
2031 n_workers = 1;
2032 g1_par_count_task.work(0);
2033 }
2034
2035 if (VerifyDuringGC) {
2036 // Verify that the counting data accumulated during marking matches
2037 // that calculated by walking the marking bitmap.
2038
2039 // Bitmaps to hold expected values
2040 BitMap expected_region_bm(_region_bm.size(), true);
2041 BitMap expected_card_bm(_card_bm.size(), true);
2042
2043 G1ParVerifyFinalCountTask g1_par_verify_task(g1h,
2044 &_region_bm,
2045 &_card_bm,
2046 &expected_region_bm,
2047 &expected_card_bm);
2048
2049 if (G1CollectedHeap::use_parallel_gc_threads()) {
2050 g1h->set_par_threads((int)n_workers);
2051 g1h->workers()->run_task(&g1_par_verify_task);
2052 // Done with the parallel phase so reset to 0.
2053 g1h->set_par_threads(0);
2054
2055 assert(g1h->check_heap_region_claim_values(HeapRegion::VerifyCountClaimValue),
2056 "sanity check");
2057 } else {
2058 g1_par_verify_task.work(0);
2059 }
2060
2061 guarantee(g1_par_verify_task.failures() == 0, "Unexpected accounting failures");
2062 }
2063
2064 size_t start_used_bytes = g1h->used();
2065 g1h->set_marking_complete();
2066
2067 double count_end = os::elapsedTime();
2068 double this_final_counting_time = (count_end - start);
2069 _total_counting_time += this_final_counting_time;
2070
2071 if (G1PrintRegionLivenessInfo) {
2072 G1PrintRegionLivenessInfoClosure cl(gclog_or_tty, "Post-Marking");
2073 _g1h->heap_region_iterate(&cl);
2074 }
2075
2076 // Install newly created mark bitMap as "prev".
2077 swapMarkBitMaps();
2078
2079 g1h->reset_gc_time_stamp();
2080
2081 // Note end of marking in all heap regions.
2082 G1ParNoteEndTask g1_par_note_end_task(g1h, &_cleanup_list);
2083 if (G1CollectedHeap::use_parallel_gc_threads()) {
2084 g1h->set_par_threads((int)n_workers);
2085 g1h->workers()->run_task(&g1_par_note_end_task);
2086 g1h->set_par_threads(0);
2087
2088 assert(g1h->check_heap_region_claim_values(HeapRegion::NoteEndClaimValue),
2089 "sanity check");
2090 } else {
2091 g1_par_note_end_task.work(0);
2092 }
2093 g1h->check_gc_time_stamps();
2094
2095 if (!cleanup_list_is_empty()) {
2096 // The cleanup list is not empty, so we'll have to process it
2097 // concurrently. Notify anyone else that might be wanting free
2098 // regions that there will be more free regions coming soon.
2099 g1h->set_free_regions_coming();
2100 }
2101
2102 // call below, since it affects the metric by which we sort the heap
2103 // regions.
2104 if (G1ScrubRemSets) {
2105 double rs_scrub_start = os::elapsedTime();
2106 G1ParScrubRemSetTask g1_par_scrub_rs_task(g1h, &_region_bm, &_card_bm);
2107 if (G1CollectedHeap::use_parallel_gc_threads()) {
2108 g1h->set_par_threads((int)n_workers);
2109 g1h->workers()->run_task(&g1_par_scrub_rs_task);
2110 g1h->set_par_threads(0);
2111
2112 assert(g1h->check_heap_region_claim_values(
2113 HeapRegion::ScrubRemSetClaimValue),
2114 "sanity check");
2115 } else {
2116 g1_par_scrub_rs_task.work(0);
2117 }
2118
2119 double rs_scrub_end = os::elapsedTime();
2120 double this_rs_scrub_time = (rs_scrub_end - rs_scrub_start);
2121 _total_rs_scrub_time += this_rs_scrub_time;
2122 }
2123
2124 // this will also free any regions totally full of garbage objects,
2125 // and sort the regions.
2126 g1h->g1_policy()->record_concurrent_mark_cleanup_end((int)n_workers);
2127
2128 // Statistics.
2129 double end = os::elapsedTime();
2130 _cleanup_times.add((end - start) * 1000.0);
2131
2132 if (G1Log::fine()) {
2133 g1h->print_size_transition(gclog_or_tty,
2134 start_used_bytes,
2135 g1h->used(),
2136 g1h->capacity());
2137 }
2138
2139 // Clean up will have freed any regions completely full of garbage.
2140 // Update the soft reference policy with the new heap occupancy.
2141 Universe::update_heap_info_at_gc();
2142
2143 // We need to make this be a "collection" so any collection pause that
2144 // races with it goes around and waits for completeCleanup to finish.
2145 g1h->increment_total_collections();
2146
2147 // We reclaimed old regions so we should calculate the sizes to make
2148 // sure we update the old gen/space data.
2149 g1h->g1mm()->update_sizes();
2150
2151 if (VerifyDuringGC) {
2152 HandleMark hm; // handle scope
2153 Universe::heap()->prepare_for_verify();
2154 Universe::verify(VerifyOption_G1UsePrevMarking,
2155 " VerifyDuringGC:(after)");
2156 }
2157 g1h->check_bitmaps("Cleanup End");
2158
2159 g1h->verify_region_sets_optional();
2160 g1h->trace_heap_after_concurrent_cycle();
2161 }
2162
2163 void ConcurrentMark::completeCleanup() {
2164 if (has_aborted()) return;
2165
2166 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2167
2168 _cleanup_list.verify_optional();
2169 FreeRegionList tmp_free_list("Tmp Free List");
2170
2171 if (G1ConcRegionFreeingVerbose) {
2172 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : "
2173 "cleanup list has %u entries",
2174 _cleanup_list.length());
2175 }
2176
2177 // Noone else should be accessing the _cleanup_list at this point,
2178 // so it's not necessary to take any locks
2179 while (!_cleanup_list.is_empty()) {
2180 HeapRegion* hr = _cleanup_list.remove_head();
2181 assert(hr != NULL, "Got NULL from a non-empty list");
2182 hr->par_clear();
2183 tmp_free_list.add_ordered(hr);
2184
2185 // Instead of adding one region at a time to the secondary_free_list,
2186 // we accumulate them in the local list and move them a few at a
2187 // time. This also cuts down on the number of notify_all() calls
2188 // we do during this process. We'll also append the local list when
2189 // _cleanup_list is empty (which means we just removed the last
2190 // region from the _cleanup_list).
2191 if ((tmp_free_list.length() % G1SecondaryFreeListAppendLength == 0) ||
2192 _cleanup_list.is_empty()) {
2193 if (G1ConcRegionFreeingVerbose) {
2194 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : "
2195 "appending %u entries to the secondary_free_list, "
2196 "cleanup list still has %u entries",
2197 tmp_free_list.length(),
2198 _cleanup_list.length());
2199 }
2200
2201 {
2202 MutexLockerEx x(SecondaryFreeList_lock, Mutex::_no_safepoint_check_flag);
2203 g1h->secondary_free_list_add(&tmp_free_list);
2204 SecondaryFreeList_lock->notify_all();
2205 }
2206
2207 if (G1StressConcRegionFreeing) {
2208 for (uintx i = 0; i < G1StressConcRegionFreeingDelayMillis; ++i) {
2209 os::sleep(Thread::current(), (jlong) 1, false);
2210 }
2211 }
2212 }
2213 }
2214 assert(tmp_free_list.is_empty(), "post-condition");
2215 }
2216
2217 // Supporting Object and Oop closures for reference discovery
2218 // and processing in during marking
2219
2220 bool G1CMIsAliveClosure::do_object_b(oop obj) {
2221 HeapWord* addr = (HeapWord*)obj;
2222 return addr != NULL &&
2223 (!_g1->is_in_g1_reserved(addr) || !_g1->is_obj_ill(obj));
2224 }
2225
2226 // 'Keep Alive' oop closure used by both serial parallel reference processing.
2227 // Uses the CMTask associated with a worker thread (for serial reference
2228 // processing the CMTask for worker 0 is used) to preserve (mark) and
2229 // trace referent objects.
2230 //
2231 // Using the CMTask and embedded local queues avoids having the worker
2232 // threads operating on the global mark stack. This reduces the risk
2233 // of overflowing the stack - which we would rather avoid at this late
2234 // state. Also using the tasks' local queues removes the potential
2235 // of the workers interfering with each other that could occur if
2236 // operating on the global stack.
2237
2238 class G1CMKeepAliveAndDrainClosure: public OopClosure {
2239 ConcurrentMark* _cm;
2240 CMTask* _task;
2241 int _ref_counter_limit;
2242 int _ref_counter;
2243 bool _is_serial;
2244 public:
2245 G1CMKeepAliveAndDrainClosure(ConcurrentMark* cm, CMTask* task, bool is_serial) :
2246 _cm(cm), _task(task), _is_serial(is_serial),
2247 _ref_counter_limit(G1RefProcDrainInterval) {
2248 assert(_ref_counter_limit > 0, "sanity");
2249 assert(!_is_serial || _task->worker_id() == 0, "only task 0 for serial code");
2250 _ref_counter = _ref_counter_limit;
2251 }
2252
2253 virtual void do_oop(narrowOop* p) { do_oop_work(p); }
2254 virtual void do_oop( oop* p) { do_oop_work(p); }
2255
2256 template <class T> void do_oop_work(T* p) {
2257 if (!_cm->has_overflown()) {
2258 oop obj = oopDesc::load_decode_heap_oop(p);
2259 if (_cm->verbose_high()) {
2260 gclog_or_tty->print_cr("\t[%u] we're looking at location "
2261 "*"PTR_FORMAT" = "PTR_FORMAT,
2262 _task->worker_id(), p2i(p), p2i((void*) obj));
2263 }
2264
2265 _task->deal_with_reference(obj);
2266 _ref_counter--;
2267
2268 if (_ref_counter == 0) {
2269 // We have dealt with _ref_counter_limit references, pushing them
2270 // and objects reachable from them on to the local stack (and
2271 // possibly the global stack). Call CMTask::do_marking_step() to
2272 // process these entries.
2273 //
2274 // We call CMTask::do_marking_step() in a loop, which we'll exit if
2275 // there's nothing more to do (i.e. we're done with the entries that
2276 // were pushed as a result of the CMTask::deal_with_reference() calls
2277 // above) or we overflow.
2278 //
2279 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted()
2280 // flag while there may still be some work to do. (See the comment at
2281 // the beginning of CMTask::do_marking_step() for those conditions -
2282 // one of which is reaching the specified time target.) It is only
2283 // when CMTask::do_marking_step() returns without setting the
2284 // has_aborted() flag that the marking step has completed.
2285 do {
2286 double mark_step_duration_ms = G1ConcMarkStepDurationMillis;
2287 _task->do_marking_step(mark_step_duration_ms,
2288 false /* do_termination */,
2289 _is_serial);
2290 } while (_task->has_aborted() && !_cm->has_overflown());
2291 _ref_counter = _ref_counter_limit;
2292 }
2293 } else {
2294 if (_cm->verbose_high()) {
2295 gclog_or_tty->print_cr("\t[%u] CM Overflow", _task->worker_id());
2296 }
2297 }
2298 }
2299 };
2300
2301 // 'Drain' oop closure used by both serial and parallel reference processing.
2302 // Uses the CMTask associated with a given worker thread (for serial
2303 // reference processing the CMtask for worker 0 is used). Calls the
2304 // do_marking_step routine, with an unbelievably large timeout value,
2305 // to drain the marking data structures of the remaining entries
2306 // added by the 'keep alive' oop closure above.
2307
2308 class G1CMDrainMarkingStackClosure: public VoidClosure {
2309 ConcurrentMark* _cm;
2310 CMTask* _task;
2311 bool _is_serial;
2312 public:
2313 G1CMDrainMarkingStackClosure(ConcurrentMark* cm, CMTask* task, bool is_serial) :
2314 _cm(cm), _task(task), _is_serial(is_serial) {
2315 assert(!_is_serial || _task->worker_id() == 0, "only task 0 for serial code");
2316 }
2317
2318 void do_void() {
2319 do {
2320 if (_cm->verbose_high()) {
2321 gclog_or_tty->print_cr("\t[%u] Drain: Calling do_marking_step - serial: %s",
2322 _task->worker_id(), BOOL_TO_STR(_is_serial));
2323 }
2324
2325 // We call CMTask::do_marking_step() to completely drain the local
2326 // and global marking stacks of entries pushed by the 'keep alive'
2327 // oop closure (an instance of G1CMKeepAliveAndDrainClosure above).
2328 //
2329 // CMTask::do_marking_step() is called in a loop, which we'll exit
2330 // if there's nothing more to do (i.e. we've completely drained the
2331 // entries that were pushed as a a result of applying the 'keep alive'
2332 // closure to the entries on the discovered ref lists) or we overflow
2333 // the global marking stack.
2334 //
2335 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted()
2336 // flag while there may still be some work to do. (See the comment at
2337 // the beginning of CMTask::do_marking_step() for those conditions -
2338 // one of which is reaching the specified time target.) It is only
2339 // when CMTask::do_marking_step() returns without setting the
2340 // has_aborted() flag that the marking step has completed.
2341
2342 _task->do_marking_step(1000000000.0 /* something very large */,
2343 true /* do_termination */,
2344 _is_serial);
2345 } while (_task->has_aborted() && !_cm->has_overflown());
2346 }
2347 };
2348
2349 // Implementation of AbstractRefProcTaskExecutor for parallel
2350 // reference processing at the end of G1 concurrent marking
2351
2352 class G1CMRefProcTaskExecutor: public AbstractRefProcTaskExecutor {
2353 private:
2354 G1CollectedHeap* _g1h;
2355 ConcurrentMark* _cm;
2356 WorkGang* _workers;
2357 int _active_workers;
2358
2359 public:
2360 G1CMRefProcTaskExecutor(G1CollectedHeap* g1h,
2361 ConcurrentMark* cm,
2362 WorkGang* workers,
2363 int n_workers) :
2364 _g1h(g1h), _cm(cm),
2365 _workers(workers), _active_workers(n_workers) { }
2366
2367 // Executes the given task using concurrent marking worker threads.
2368 virtual void execute(ProcessTask& task);
2369 virtual void execute(EnqueueTask& task);
2370 };
2371
2372 class G1CMRefProcTaskProxy: public AbstractGangTask {
2373 typedef AbstractRefProcTaskExecutor::ProcessTask ProcessTask;
2374 ProcessTask& _proc_task;
2375 G1CollectedHeap* _g1h;
2376 ConcurrentMark* _cm;
2377
2378 public:
2379 G1CMRefProcTaskProxy(ProcessTask& proc_task,
2380 G1CollectedHeap* g1h,
2381 ConcurrentMark* cm) :
2382 AbstractGangTask("Process reference objects in parallel"),
2383 _proc_task(proc_task), _g1h(g1h), _cm(cm) {
2384 ReferenceProcessor* rp = _g1h->ref_processor_cm();
2385 assert(rp->processing_is_mt(), "shouldn't be here otherwise");
2386 }
2387
2388 virtual void work(uint worker_id) {
2389 CMTask* task = _cm->task(worker_id);
2390 G1CMIsAliveClosure g1_is_alive(_g1h);
2391 G1CMKeepAliveAndDrainClosure g1_par_keep_alive(_cm, task, false /* is_serial */);
2392 G1CMDrainMarkingStackClosure g1_par_drain(_cm, task, false /* is_serial */);
2393
2394 _proc_task.work(worker_id, g1_is_alive, g1_par_keep_alive, g1_par_drain);
2395 }
2396 };
2397
2398 void G1CMRefProcTaskExecutor::execute(ProcessTask& proc_task) {
2399 assert(_workers != NULL, "Need parallel worker threads.");
2400 assert(_g1h->ref_processor_cm()->processing_is_mt(), "processing is not MT");
2401
2402 G1CMRefProcTaskProxy proc_task_proxy(proc_task, _g1h, _cm);
2403
2404 // We need to reset the concurrency level before each
2405 // proxy task execution, so that the termination protocol
2406 // and overflow handling in CMTask::do_marking_step() knows
2407 // how many workers to wait for.
2408 _cm->set_concurrency(_active_workers);
2409 _g1h->set_par_threads(_active_workers);
2410 _workers->run_task(&proc_task_proxy);
2411 _g1h->set_par_threads(0);
2412 }
2413
2414 class G1CMRefEnqueueTaskProxy: public AbstractGangTask {
2415 typedef AbstractRefProcTaskExecutor::EnqueueTask EnqueueTask;
2416 EnqueueTask& _enq_task;
2417
2418 public:
2419 G1CMRefEnqueueTaskProxy(EnqueueTask& enq_task) :
2420 AbstractGangTask("Enqueue reference objects in parallel"),
2421 _enq_task(enq_task) { }
2422
2423 virtual void work(uint worker_id) {
2424 _enq_task.work(worker_id);
2425 }
2426 };
2427
2428 void G1CMRefProcTaskExecutor::execute(EnqueueTask& enq_task) {
2429 assert(_workers != NULL, "Need parallel worker threads.");
2430 assert(_g1h->ref_processor_cm()->processing_is_mt(), "processing is not MT");
2431
2432 G1CMRefEnqueueTaskProxy enq_task_proxy(enq_task);
2433
2434 // Not strictly necessary but...
2435 //
2436 // We need to reset the concurrency level before each
2437 // proxy task execution, so that the termination protocol
2438 // and overflow handling in CMTask::do_marking_step() knows
2439 // how many workers to wait for.
2440 _cm->set_concurrency(_active_workers);
2441 _g1h->set_par_threads(_active_workers);
2442 _workers->run_task(&enq_task_proxy);
2443 _g1h->set_par_threads(0);
2444 }
2445
2446 void ConcurrentMark::weakRefsWork(bool clear_all_soft_refs) {
2447 if (has_overflown()) {
2448 // Skip processing the discovered references if we have
2449 // overflown the global marking stack. Reference objects
2450 // only get discovered once so it is OK to not
2451 // de-populate the discovered reference lists. We could have,
2452 // but the only benefit would be that, when marking restarts,
2453 // less reference objects are discovered.
2454 return;
2455 }
2456
2457 ResourceMark rm;
2458 HandleMark hm;
2459
2460 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2461
2462 // Is alive closure.
2463 G1CMIsAliveClosure g1_is_alive(g1h);
2464
2465 // Inner scope to exclude the cleaning of the string and symbol
2466 // tables from the displayed time.
2467 {
2468 if (G1Log::finer()) {
2469 gclog_or_tty->put(' ');
2470 }
2471 GCTraceTime t("GC ref-proc", G1Log::finer(), false, g1h->gc_timer_cm());
2472
2473 ReferenceProcessor* rp = g1h->ref_processor_cm();
2474
2475 // See the comment in G1CollectedHeap::ref_processing_init()
2476 // about how reference processing currently works in G1.
2477
2478 // Set the soft reference policy
2479 rp->setup_policy(clear_all_soft_refs);
2480 assert(_markStack.isEmpty(), "mark stack should be empty");
2481
2482 // Instances of the 'Keep Alive' and 'Complete GC' closures used
2483 // in serial reference processing. Note these closures are also
2484 // used for serially processing (by the the current thread) the
2485 // JNI references during parallel reference processing.
2486 //
2487 // These closures do not need to synchronize with the worker
2488 // threads involved in parallel reference processing as these
2489 // instances are executed serially by the current thread (e.g.
2490 // reference processing is not multi-threaded and is thus
2491 // performed by the current thread instead of a gang worker).
2492 //
2493 // The gang tasks involved in parallel reference processing create
2494 // their own instances of these closures, which do their own
2495 // synchronization among themselves.
2496 G1CMKeepAliveAndDrainClosure g1_keep_alive(this, task(0), true /* is_serial */);
2497 G1CMDrainMarkingStackClosure g1_drain_mark_stack(this, task(0), true /* is_serial */);
2498
2499 // We need at least one active thread. If reference processing
2500 // is not multi-threaded we use the current (VMThread) thread,
2501 // otherwise we use the work gang from the G1CollectedHeap and
2502 // we utilize all the worker threads we can.
2503 bool processing_is_mt = rp->processing_is_mt() && g1h->workers() != NULL;
2504 uint active_workers = (processing_is_mt ? g1h->workers()->active_workers() : 1U);
2505 active_workers = MAX2(MIN2(active_workers, _max_worker_id), 1U);
2506
2507 // Parallel processing task executor.
2508 G1CMRefProcTaskExecutor par_task_executor(g1h, this,
2509 g1h->workers(), active_workers);
2510 AbstractRefProcTaskExecutor* executor = (processing_is_mt ? &par_task_executor : NULL);
2511
2512 // Set the concurrency level. The phase was already set prior to
2513 // executing the remark task.
2514 set_concurrency(active_workers);
2515
2516 // Set the degree of MT processing here. If the discovery was done MT,
2517 // the number of threads involved during discovery could differ from
2518 // the number of active workers. This is OK as long as the discovered
2519 // Reference lists are balanced (see balance_all_queues() and balance_queues()).
2520 rp->set_active_mt_degree(active_workers);
2521
2522 // Process the weak references.
2523 const ReferenceProcessorStats& stats =
2524 rp->process_discovered_references(&g1_is_alive,
2525 &g1_keep_alive,
2526 &g1_drain_mark_stack,
2527 executor,
2528 g1h->gc_timer_cm());
2529 g1h->gc_tracer_cm()->report_gc_reference_stats(stats);
2530
2531 // The do_oop work routines of the keep_alive and drain_marking_stack
2532 // oop closures will set the has_overflown flag if we overflow the
2533 // global marking stack.
2534
2535 assert(_markStack.overflow() || _markStack.isEmpty(),
2536 "mark stack should be empty (unless it overflowed)");
2537
2538 if (_markStack.overflow()) {
2539 // This should have been done already when we tried to push an
2540 // entry on to the global mark stack. But let's do it again.
2541 set_has_overflown();
2542 }
2543
2544 assert(rp->num_q() == active_workers, "why not");
2545
2546 rp->enqueue_discovered_references(executor);
2547
2548 rp->verify_no_references_recorded();
2549 assert(!rp->discovery_enabled(), "Post condition");
2550 }
2551
2552 if (has_overflown()) {
2553 // We can not trust g1_is_alive if the marking stack overflowed
2554 return;
2555 }
2556
2557 g1h->unlink_string_and_symbol_table(&g1_is_alive,
2558 /* process_strings */ false, // currently strings are always roots
2559 /* process_symbols */ true);
2560 }
2561
2562 void ConcurrentMark::swapMarkBitMaps() {
2563 CMBitMapRO* temp = _prevMarkBitMap;
2564 _prevMarkBitMap = (CMBitMapRO*)_nextMarkBitMap;
2565 _nextMarkBitMap = (CMBitMap*) temp;
2566 }
2567
2568 class CMRemarkTask: public AbstractGangTask {
2569 private:
2570 ConcurrentMark* _cm;
2571 bool _is_serial;
2572 public:
2573 void work(uint worker_id) {
2574 // Since all available tasks are actually started, we should
2575 // only proceed if we're supposed to be active.
2576 if (worker_id < _cm->active_tasks()) {
2577 CMTask* task = _cm->task(worker_id);
2578 task->record_start_time();
2579 do {
2580 task->do_marking_step(1000000000.0 /* something very large */,
2581 true /* do_termination */,
2582 _is_serial);
2583 } while (task->has_aborted() && !_cm->has_overflown());
2584 // If we overflow, then we do not want to restart. We instead
2585 // want to abort remark and do concurrent marking again.
2586 task->record_end_time();
2587 }
2588 }
2589
2590 CMRemarkTask(ConcurrentMark* cm, int active_workers, bool is_serial) :
2591 AbstractGangTask("Par Remark"), _cm(cm), _is_serial(is_serial) {
2592 _cm->terminator()->reset_for_reuse(active_workers);
2593 }
2594 };
2595
2596 void ConcurrentMark::checkpointRootsFinalWork() {
2597 ResourceMark rm;
2598 HandleMark hm;
2599 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2600
2601 g1h->ensure_parsability(false);
2602
2603 if (G1CollectedHeap::use_parallel_gc_threads()) {
2604 G1CollectedHeap::StrongRootsScope srs(g1h);
2605 // this is remark, so we'll use up all active threads
2606 uint active_workers = g1h->workers()->active_workers();
2607 if (active_workers == 0) {
2608 assert(active_workers > 0, "Should have been set earlier");
2609 active_workers = (uint) ParallelGCThreads;
2610 g1h->workers()->set_active_workers(active_workers);
2611 }
2612 set_concurrency_and_phase(active_workers, false /* concurrent */);
2613 // Leave _parallel_marking_threads at it's
2614 // value originally calculated in the ConcurrentMark
2615 // constructor and pass values of the active workers
2616 // through the gang in the task.
2617
2618 CMRemarkTask remarkTask(this, active_workers, false /* is_serial */);
2619 // We will start all available threads, even if we decide that the
2620 // active_workers will be fewer. The extra ones will just bail out
2621 // immediately.
2622 g1h->set_par_threads(active_workers);
2623 g1h->workers()->run_task(&remarkTask);
2624 g1h->set_par_threads(0);
2625 } else {
2626 G1CollectedHeap::StrongRootsScope srs(g1h);
2627 uint active_workers = 1;
2628 set_concurrency_and_phase(active_workers, false /* concurrent */);
2629
2630 // Note - if there's no work gang then the VMThread will be
2631 // the thread to execute the remark - serially. We have
2632 // to pass true for the is_serial parameter so that
2633 // CMTask::do_marking_step() doesn't enter the sync
2634 // barriers in the event of an overflow. Doing so will
2635 // cause an assert that the current thread is not a
2636 // concurrent GC thread.
2637 CMRemarkTask remarkTask(this, active_workers, true /* is_serial*/);
2638 remarkTask.work(0);
2639 }
2640 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
2641 guarantee(has_overflown() ||
2642 satb_mq_set.completed_buffers_num() == 0,
2643 err_msg("Invariant: has_overflown = %s, num buffers = %d",
2644 BOOL_TO_STR(has_overflown()),
2645 satb_mq_set.completed_buffers_num()));
2646
2647 print_stats();
2648 }
2649
2650 #ifndef PRODUCT
2651
2652 class PrintReachableOopClosure: public OopClosure {
2653 private:
2654 G1CollectedHeap* _g1h;
2655 outputStream* _out;
2656 VerifyOption _vo;
2657 bool _all;
2658
2659 public:
2660 PrintReachableOopClosure(outputStream* out,
2661 VerifyOption vo,
2662 bool all) :
2663 _g1h(G1CollectedHeap::heap()),
2664 _out(out), _vo(vo), _all(all) { }
2665
2666 void do_oop(narrowOop* p) { do_oop_work(p); }
2667 void do_oop( oop* p) { do_oop_work(p); }
2668
2669 template <class T> void do_oop_work(T* p) {
2670 oop obj = oopDesc::load_decode_heap_oop(p);
2671 const char* str = NULL;
2672 const char* str2 = "";
2673
2674 if (obj == NULL) {
2675 str = "";
2676 } else if (!_g1h->is_in_g1_reserved(obj)) {
2677 str = " O";
2678 } else {
2679 HeapRegion* hr = _g1h->heap_region_containing(obj);
2680 bool over_tams = _g1h->allocated_since_marking(obj, hr, _vo);
2681 bool marked = _g1h->is_marked(obj, _vo);
2682
2683 if (over_tams) {
2684 str = " >";
2685 if (marked) {
2686 str2 = " AND MARKED";
2687 }
2688 } else if (marked) {
2689 str = " M";
2690 } else {
2691 str = " NOT";
2692 }
2693 }
2694
2695 _out->print_cr(" "PTR_FORMAT": "PTR_FORMAT"%s%s",
2696 p2i(p), p2i((void*) obj), str, str2);
2697 }
2698 };
2699
2700 class PrintReachableObjectClosure : public ObjectClosure {
2701 private:
2702 G1CollectedHeap* _g1h;
2703 outputStream* _out;
2704 VerifyOption _vo;
2705 bool _all;
2706 HeapRegion* _hr;
2707
2708 public:
2709 PrintReachableObjectClosure(outputStream* out,
2710 VerifyOption vo,
2711 bool all,
2712 HeapRegion* hr) :
2713 _g1h(G1CollectedHeap::heap()),
2714 _out(out), _vo(vo), _all(all), _hr(hr) { }
2715
2716 void do_object(oop o) {
2717 bool over_tams = _g1h->allocated_since_marking(o, _hr, _vo);
2718 bool marked = _g1h->is_marked(o, _vo);
2719 bool print_it = _all || over_tams || marked;
2720
2721 if (print_it) {
2722 _out->print_cr(" "PTR_FORMAT"%s",
2723 p2i((void *)o), (over_tams) ? " >" : (marked) ? " M" : "");
2724 PrintReachableOopClosure oopCl(_out, _vo, _all);
2725 o->oop_iterate_no_header(&oopCl);
2726 }
2727 }
2728 };
2729
2730 class PrintReachableRegionClosure : public HeapRegionClosure {
2731 private:
2732 G1CollectedHeap* _g1h;
2733 outputStream* _out;
2734 VerifyOption _vo;
2735 bool _all;
2736
2737 public:
2738 bool doHeapRegion(HeapRegion* hr) {
2739 HeapWord* b = hr->bottom();
2740 HeapWord* e = hr->end();
2741 HeapWord* t = hr->top();
2742 HeapWord* p = _g1h->top_at_mark_start(hr, _vo);
2743 _out->print_cr("** ["PTR_FORMAT", "PTR_FORMAT"] top: "PTR_FORMAT" "
2744 "TAMS: " PTR_FORMAT, p2i(b), p2i(e), p2i(t), p2i(p));
2745 _out->cr();
2746
2747 HeapWord* from = b;
2748 HeapWord* to = t;
2749
2750 if (to > from) {
2751 _out->print_cr("Objects in [" PTR_FORMAT ", " PTR_FORMAT "]", p2i(from), p2i(to));
2752 _out->cr();
2753 PrintReachableObjectClosure ocl(_out, _vo, _all, hr);
2754 hr->object_iterate_mem_careful(MemRegion(from, to), &ocl);
2755 _out->cr();
2756 }
2757
2758 return false;
2759 }
2760
2761 PrintReachableRegionClosure(outputStream* out,
2762 VerifyOption vo,
2763 bool all) :
2764 _g1h(G1CollectedHeap::heap()), _out(out), _vo(vo), _all(all) { }
2765 };
2766
2767 void ConcurrentMark::print_reachable(const char* str,
2768 VerifyOption vo,
2769 bool all) {
2770 gclog_or_tty->cr();
2771 gclog_or_tty->print_cr("== Doing heap dump... ");
2772
2773 if (G1PrintReachableBaseFile == NULL) {
2774 gclog_or_tty->print_cr(" #### error: no base file defined");
2775 return;
2776 }
2777
2778 if (strlen(G1PrintReachableBaseFile) + 1 + strlen(str) >
2779 (JVM_MAXPATHLEN - 1)) {
2780 gclog_or_tty->print_cr(" #### error: file name too long");
2781 return;
2782 }
2783
2784 char file_name[JVM_MAXPATHLEN];
2785 sprintf(file_name, "%s.%s", G1PrintReachableBaseFile, str);
2786 gclog_or_tty->print_cr(" dumping to file %s", file_name);
2787
2788 fileStream fout(file_name);
2789 if (!fout.is_open()) {
2790 gclog_or_tty->print_cr(" #### error: could not open file");
2791 return;
2792 }
2793
2794 outputStream* out = &fout;
2795 out->print_cr("-- USING %s", _g1h->top_at_mark_start_str(vo));
2796 out->cr();
2797
2798 out->print_cr("--- ITERATING OVER REGIONS");
2799 out->cr();
2800 PrintReachableRegionClosure rcl(out, vo, all);
2801 _g1h->heap_region_iterate(&rcl);
2802 out->cr();
2803
2804 gclog_or_tty->print_cr(" done");
2805 gclog_or_tty->flush();
2806 }
2807
2808 #endif // PRODUCT
2809
2810 void ConcurrentMark::clearRangePrevBitmap(MemRegion mr) {
2811 // Note we are overriding the read-only view of the prev map here, via
2812 // the cast.
2813 ((CMBitMap*)_prevMarkBitMap)->clearRange(mr);
2814 }
2815
2816 void ConcurrentMark::clearRangeNextBitmap(MemRegion mr) {
2817 _nextMarkBitMap->clearRange(mr);
2818 }
2819
2820 void ConcurrentMark::clearRangeBothBitmaps(MemRegion mr) {
2821 clearRangePrevBitmap(mr);
2822 clearRangeNextBitmap(mr);
2823 }
2824
2825 HeapRegion*
2826 ConcurrentMark::claim_region(uint worker_id) {
2827 // "checkpoint" the finger
2828 HeapWord* finger = _finger;
2829
2830 // _heap_end will not change underneath our feet; it only changes at
2831 // yield points.
2832 while (finger < _heap_end) {
2833 assert(_g1h->is_in_g1_reserved(finger), "invariant");
2834
2835 // Note on how this code handles humongous regions. In the
2836 // normal case the finger will reach the start of a "starts
2837 // humongous" (SH) region. Its end will either be the end of the
2838 // last "continues humongous" (CH) region in the sequence, or the
2839 // standard end of the SH region (if the SH is the only region in
2840 // the sequence). That way claim_region() will skip over the CH
2841 // regions. However, there is a subtle race between a CM thread
2842 // executing this method and a mutator thread doing a humongous
2843 // object allocation. The two are not mutually exclusive as the CM
2844 // thread does not need to hold the Heap_lock when it gets
2845 // here. So there is a chance that claim_region() will come across
2846 // a free region that's in the progress of becoming a SH or a CH
2847 // region. In the former case, it will either
2848 // a) Miss the update to the region's end, in which case it will
2849 // visit every subsequent CH region, will find their bitmaps
2850 // empty, and do nothing, or
2851 // b) Will observe the update of the region's end (in which case
2852 // it will skip the subsequent CH regions).
2853 // If it comes across a region that suddenly becomes CH, the
2854 // scenario will be similar to b). So, the race between
2855 // claim_region() and a humongous object allocation might force us
2856 // to do a bit of unnecessary work (due to some unnecessary bitmap
2857 // iterations) but it should not introduce and correctness issues.
2858 HeapRegion* curr_region = _g1h->heap_region_containing_raw(finger);
2859 HeapWord* bottom = curr_region->bottom();
2860 HeapWord* end = curr_region->end();
2861 HeapWord* limit = curr_region->next_top_at_mark_start();
2862
2863 if (verbose_low()) {
2864 gclog_or_tty->print_cr("[%u] curr_region = "PTR_FORMAT" "
2865 "["PTR_FORMAT", "PTR_FORMAT"), "
2866 "limit = "PTR_FORMAT,
2867 worker_id, p2i(curr_region), p2i(bottom), p2i(end), p2i(limit));
2868 }
2869
2870 // Is the gap between reading the finger and doing the CAS too long?
2871 HeapWord* res = (HeapWord*) Atomic::cmpxchg_ptr(end, &_finger, finger);
2872 if (res == finger) {
2873 // we succeeded
2874
2875 // notice that _finger == end cannot be guaranteed here since,
2876 // someone else might have moved the finger even further
2877 assert(_finger >= end, "the finger should have moved forward");
2878
2879 if (verbose_low()) {
2880 gclog_or_tty->print_cr("[%u] we were successful with region = "
2881 PTR_FORMAT, worker_id, p2i(curr_region));
2882 }
2883
2884 if (limit > bottom) {
2885 if (verbose_low()) {
2886 gclog_or_tty->print_cr("[%u] region "PTR_FORMAT" is not empty, "
2887 "returning it ", worker_id, p2i(curr_region));
2888 }
2889 return curr_region;
2890 } else {
2891 assert(limit == bottom,
2892 "the region limit should be at bottom");
2893 if (verbose_low()) {
2894 gclog_or_tty->print_cr("[%u] region "PTR_FORMAT" is empty, "
2895 "returning NULL", worker_id, p2i(curr_region));
2896 }
2897 // we return NULL and the caller should try calling
2898 // claim_region() again.
2899 return NULL;
2900 }
2901 } else {
2902 assert(_finger > finger, "the finger should have moved forward");
2903 if (verbose_low()) {
2904 gclog_or_tty->print_cr("[%u] somebody else moved the finger, "
2905 "global finger = "PTR_FORMAT", "
2906 "our finger = "PTR_FORMAT,
2907 worker_id, p2i(_finger), p2i(finger));
2908 }
2909
2910 // read it again
2911 finger = _finger;
2912 }
2913 }
2914
2915 return NULL;
2916 }
2917
2918 #ifndef PRODUCT
2919 enum VerifyNoCSetOopsPhase {
2920 VerifyNoCSetOopsStack,
2921 VerifyNoCSetOopsQueues,
2922 VerifyNoCSetOopsSATBCompleted,
2923 VerifyNoCSetOopsSATBThread
2924 };
2925
2926 class VerifyNoCSetOopsClosure : public OopClosure, public ObjectClosure {
2927 private:
2928 G1CollectedHeap* _g1h;
2929 VerifyNoCSetOopsPhase _phase;
2930 int _info;
2931
2932 const char* phase_str() {
2933 switch (_phase) {
2934 case VerifyNoCSetOopsStack: return "Stack";
2935 case VerifyNoCSetOopsQueues: return "Queue";
2936 case VerifyNoCSetOopsSATBCompleted: return "Completed SATB Buffers";
2937 case VerifyNoCSetOopsSATBThread: return "Thread SATB Buffers";
2938 default: ShouldNotReachHere();
2939 }
2940 return NULL;
2941 }
2942
2943 void do_object_work(oop obj) {
2944 guarantee(!_g1h->obj_in_cs(obj),
2945 err_msg("obj: "PTR_FORMAT" in CSet, phase: %s, info: %d",
2946 p2i((void*) obj), phase_str(), _info));
2947 }
2948
2949 public:
2950 VerifyNoCSetOopsClosure() : _g1h(G1CollectedHeap::heap()) { }
2951
2952 void set_phase(VerifyNoCSetOopsPhase phase, int info = -1) {
2953 _phase = phase;
2954 _info = info;
2955 }
2956
2957 virtual void do_oop(oop* p) {
2958 oop obj = oopDesc::load_decode_heap_oop(p);
2959 do_object_work(obj);
2960 }
2961
2962 virtual void do_oop(narrowOop* p) {
2963 // We should not come across narrow oops while scanning marking
2964 // stacks and SATB buffers.
2965 ShouldNotReachHere();
2966 }
2967
2968 virtual void do_object(oop obj) {
2969 do_object_work(obj);
2970 }
2971 };
2972
2973 void ConcurrentMark::verify_no_cset_oops(bool verify_stacks,
2974 bool verify_enqueued_buffers,
2975 bool verify_thread_buffers,
2976 bool verify_fingers) {
2977 assert(SafepointSynchronize::is_at_safepoint(), "should be at a safepoint");
2978 if (!G1CollectedHeap::heap()->mark_in_progress()) {
2979 return;
2980 }
2981
2982 VerifyNoCSetOopsClosure cl;
2983
2984 if (verify_stacks) {
2985 // Verify entries on the global mark stack
2986 cl.set_phase(VerifyNoCSetOopsStack);
2987 _markStack.oops_do(&cl);
2988
2989 // Verify entries on the task queues
2990 for (uint i = 0; i < _max_worker_id; i += 1) {
2991 cl.set_phase(VerifyNoCSetOopsQueues, i);
2992 CMTaskQueue* queue = _task_queues->queue(i);
2993 queue->oops_do(&cl);
2994 }
2995 }
2996
2997 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set();
2998
2999 // Verify entries on the enqueued SATB buffers
3000 if (verify_enqueued_buffers) {
3001 cl.set_phase(VerifyNoCSetOopsSATBCompleted);
3002 satb_qs.iterate_completed_buffers_read_only(&cl);
3003 }
3004
3005 // Verify entries on the per-thread SATB buffers
3006 if (verify_thread_buffers) {
3007 cl.set_phase(VerifyNoCSetOopsSATBThread);
3008 satb_qs.iterate_thread_buffers_read_only(&cl);
3009 }
3010
3011 if (verify_fingers) {
3012 // Verify the global finger
3013 HeapWord* global_finger = finger();
3014 if (global_finger != NULL && global_finger < _heap_end) {
3015 // The global finger always points to a heap region boundary. We
3016 // use heap_region_containing_raw() to get the containing region
3017 // given that the global finger could be pointing to a free region
3018 // which subsequently becomes continues humongous. If that
3019 // happens, heap_region_containing() will return the bottom of the
3020 // corresponding starts humongous region and the check below will
3021 // not hold any more.
3022 HeapRegion* global_hr = _g1h->heap_region_containing_raw(global_finger);
3023 guarantee(global_finger == global_hr->bottom(),
3024 err_msg("global finger: "PTR_FORMAT" region: "HR_FORMAT,
3025 p2i(global_finger), HR_FORMAT_PARAMS(global_hr)));
3026 }
3027
3028 // Verify the task fingers
3029 assert(parallel_marking_threads() <= _max_worker_id, "sanity");
3030 for (int i = 0; i < (int) parallel_marking_threads(); i += 1) {
3031 CMTask* task = _tasks[i];
3032 HeapWord* task_finger = task->finger();
3033 if (task_finger != NULL && task_finger < _heap_end) {
3034 // See above note on the global finger verification.
3035 HeapRegion* task_hr = _g1h->heap_region_containing_raw(task_finger);
3036 guarantee(task_finger == task_hr->bottom() ||
3037 !task_hr->in_collection_set(),
3038 err_msg("task finger: "PTR_FORMAT" region: "HR_FORMAT,
3039 p2i(task_finger), HR_FORMAT_PARAMS(task_hr)));
3040 }
3041 }
3042 }
3043 }
3044 #endif // PRODUCT
3045
3046 // Aggregate the counting data that was constructed concurrently
3047 // with marking.
3048 class AggregateCountDataHRClosure: public HeapRegionClosure {
3049 G1CollectedHeap* _g1h;
3050 ConcurrentMark* _cm;
3051 CardTableModRefBS* _ct_bs;
3052 BitMap* _cm_card_bm;
3053 uint _max_worker_id;
3054
3055 public:
3056 AggregateCountDataHRClosure(G1CollectedHeap* g1h,
3057 BitMap* cm_card_bm,
3058 uint max_worker_id) :
3059 _g1h(g1h), _cm(g1h->concurrent_mark()),
3060 _ct_bs((CardTableModRefBS*) (g1h->barrier_set())),
3061 _cm_card_bm(cm_card_bm), _max_worker_id(max_worker_id) { }
3062
3063 bool doHeapRegion(HeapRegion* hr) {
3064 if (hr->continuesHumongous()) {
3065 // We will ignore these here and process them when their
3066 // associated "starts humongous" region is processed.
3067 // Note that we cannot rely on their associated
3068 // "starts humongous" region to have their bit set to 1
3069 // since, due to the region chunking in the parallel region
3070 // iteration, a "continues humongous" region might be visited
3071 // before its associated "starts humongous".
3072 return false;
3073 }
3074
3075 HeapWord* start = hr->bottom();
3076 HeapWord* limit = hr->next_top_at_mark_start();
3077 HeapWord* end = hr->end();
3078
3079 assert(start <= limit && limit <= hr->top() && hr->top() <= hr->end(),
3080 err_msg("Preconditions not met - "
3081 "start: "PTR_FORMAT", limit: "PTR_FORMAT", "
3082 "top: "PTR_FORMAT", end: "PTR_FORMAT,
3083 p2i(start), p2i(limit), p2i(hr->top()), p2i(hr->end())));
3084
3085 assert(hr->next_marked_bytes() == 0, "Precondition");
3086
3087 if (start == limit) {
3088 // NTAMS of this region has not been set so nothing to do.
3089 return false;
3090 }
3091
3092 // 'start' should be in the heap.
3093 assert(_g1h->is_in_g1_reserved(start) && _ct_bs->is_card_aligned(start), "sanity");
3094 // 'end' *may* be just beyond the end of the heap (if hr is the last region)
3095 assert(!_g1h->is_in_g1_reserved(end) || _ct_bs->is_card_aligned(end), "sanity");
3096
3097 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(start);
3098 BitMap::idx_t limit_idx = _cm->card_bitmap_index_for(limit);
3099 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(end);
3100
3101 // If ntams is not card aligned then we bump card bitmap index
3102 // for limit so that we get the all the cards spanned by
3103 // the object ending at ntams.
3104 // Note: if this is the last region in the heap then ntams
3105 // could be actually just beyond the end of the the heap;
3106 // limit_idx will then correspond to a (non-existent) card
3107 // that is also outside the heap.
3108 if (_g1h->is_in_g1_reserved(limit) && !_ct_bs->is_card_aligned(limit)) {
3109 limit_idx += 1;
3110 }
3111
3112 assert(limit_idx <= end_idx, "or else use atomics");
3113
3114 // Aggregate the "stripe" in the count data associated with hr.
3115 uint hrs_index = hr->hrs_index();
3116 size_t marked_bytes = 0;
3117
3118 for (uint i = 0; i < _max_worker_id; i += 1) {
3119 size_t* marked_bytes_array = _cm->count_marked_bytes_array_for(i);
3120 BitMap* task_card_bm = _cm->count_card_bitmap_for(i);
3121
3122 // Fetch the marked_bytes in this region for task i and
3123 // add it to the running total for this region.
3124 marked_bytes += marked_bytes_array[hrs_index];
3125
3126 // Now union the bitmaps[0,max_worker_id)[start_idx..limit_idx)
3127 // into the global card bitmap.
3128 BitMap::idx_t scan_idx = task_card_bm->get_next_one_offset(start_idx, limit_idx);
3129
3130 while (scan_idx < limit_idx) {
3131 assert(task_card_bm->at(scan_idx) == true, "should be");
3132 _cm_card_bm->set_bit(scan_idx);
3133 assert(_cm_card_bm->at(scan_idx) == true, "should be");
3134
3135 // BitMap::get_next_one_offset() can handle the case when
3136 // its left_offset parameter is greater than its right_offset
3137 // parameter. It does, however, have an early exit if
3138 // left_offset == right_offset. So let's limit the value
3139 // passed in for left offset here.
3140 BitMap::idx_t next_idx = MIN2(scan_idx + 1, limit_idx);
3141 scan_idx = task_card_bm->get_next_one_offset(next_idx, limit_idx);
3142 }
3143 }
3144
3145 // Update the marked bytes for this region.
3146 hr->add_to_marked_bytes(marked_bytes);
3147
3148 // Next heap region
3149 return false;
3150 }
3151 };
3152
3153 class G1AggregateCountDataTask: public AbstractGangTask {
3154 protected:
3155 G1CollectedHeap* _g1h;
3156 ConcurrentMark* _cm;
3157 BitMap* _cm_card_bm;
3158 uint _max_worker_id;
3159 int _active_workers;
3160
3161 public:
3162 G1AggregateCountDataTask(G1CollectedHeap* g1h,
3163 ConcurrentMark* cm,
3164 BitMap* cm_card_bm,
3165 uint max_worker_id,
3166 int n_workers) :
3167 AbstractGangTask("Count Aggregation"),
3168 _g1h(g1h), _cm(cm), _cm_card_bm(cm_card_bm),
3169 _max_worker_id(max_worker_id),
3170 _active_workers(n_workers) { }
3171
3172 void work(uint worker_id) {
3173 AggregateCountDataHRClosure cl(_g1h, _cm_card_bm, _max_worker_id);
3174
3175 if (G1CollectedHeap::use_parallel_gc_threads()) {
3176 _g1h->heap_region_par_iterate_chunked(&cl, worker_id,
3177 _active_workers,
3178 HeapRegion::AggregateCountClaimValue);
3179 } else {
3180 _g1h->heap_region_iterate(&cl);
3181 }
3182 }
3183 };
3184
3185
3186 void ConcurrentMark::aggregate_count_data() {
3187 int n_workers = (G1CollectedHeap::use_parallel_gc_threads() ?
3188 _g1h->workers()->active_workers() :
3189 1);
3190
3191 G1AggregateCountDataTask g1_par_agg_task(_g1h, this, &_card_bm,
3192 _max_worker_id, n_workers);
3193
3194 if (G1CollectedHeap::use_parallel_gc_threads()) {
3195 assert(_g1h->check_heap_region_claim_values(HeapRegion::InitialClaimValue),
3196 "sanity check");
3197 _g1h->set_par_threads(n_workers);
3198 _g1h->workers()->run_task(&g1_par_agg_task);
3199 _g1h->set_par_threads(0);
3200
3201 assert(_g1h->check_heap_region_claim_values(HeapRegion::AggregateCountClaimValue),
3202 "sanity check");
3203 _g1h->reset_heap_region_claim_values();
3204 } else {
3205 g1_par_agg_task.work(0);
3206 }
3207 }
3208
3209 // Clear the per-worker arrays used to store the per-region counting data
3210 void ConcurrentMark::clear_all_count_data() {
3211 // Clear the global card bitmap - it will be filled during
3212 // liveness count aggregation (during remark) and the
3213 // final counting task.
3214 _card_bm.clear();
3215
3216 // Clear the global region bitmap - it will be filled as part
3217 // of the final counting task.
3218 _region_bm.clear();
3219
3220 uint max_regions = _g1h->max_regions();
3221 assert(_max_worker_id > 0, "uninitialized");
3222
3223 for (uint i = 0; i < _max_worker_id; i += 1) {
3224 BitMap* task_card_bm = count_card_bitmap_for(i);
3225 size_t* marked_bytes_array = count_marked_bytes_array_for(i);
3226
3227 assert(task_card_bm->size() == _card_bm.size(), "size mismatch");
3228 assert(marked_bytes_array != NULL, "uninitialized");
3229
3230 memset(marked_bytes_array, 0, (size_t) max_regions * sizeof(size_t));
3231 task_card_bm->clear();
3232 }
3233 }
3234
3235 void ConcurrentMark::print_stats() {
3236 if (verbose_stats()) {
3237 gclog_or_tty->print_cr("---------------------------------------------------------------------");
3238 for (size_t i = 0; i < _active_tasks; ++i) {
3239 _tasks[i]->print_stats();
3240 gclog_or_tty->print_cr("---------------------------------------------------------------------");
3241 }
3242 }
3243 }
3244
3245 // abandon current marking iteration due to a Full GC
3246 void ConcurrentMark::abort() {
3247 // Clear all marks to force marking thread to do nothing
3248 _nextMarkBitMap->clearAll();
3249
3250 // Note we cannot clear the previous marking bitmap here
3251 // since VerifyDuringGC verifies the objects marked during
3252 // a full GC against the previous bitmap.
3253
3254 // Clear the liveness counting data
3255 clear_all_count_data();
3256 // Empty mark stack
3257 reset_marking_state();
3258 for (uint i = 0; i < _max_worker_id; ++i) {
3259 _tasks[i]->clear_region_fields();
3260 }
3261 _first_overflow_barrier_sync.abort();
3262 _second_overflow_barrier_sync.abort();
3263 _has_aborted = true;
3264
3265 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3266 satb_mq_set.abandon_partial_marking();
3267 // This can be called either during or outside marking, we'll read
3268 // the expected_active value from the SATB queue set.
3269 satb_mq_set.set_active_all_threads(
3270 false, /* new active value */
3271 satb_mq_set.is_active() /* expected_active */);
3272
3273 _g1h->trace_heap_after_concurrent_cycle();
3274 _g1h->register_concurrent_cycle_end();
3275 }
3276
3277 static void print_ms_time_info(const char* prefix, const char* name,
3278 NumberSeq& ns) {
3279 gclog_or_tty->print_cr("%s%5d %12s: total time = %8.2f s (avg = %8.2f ms).",
3280 prefix, ns.num(), name, ns.sum()/1000.0, ns.avg());
3281 if (ns.num() > 0) {
3282 gclog_or_tty->print_cr("%s [std. dev = %8.2f ms, max = %8.2f ms]",
3283 prefix, ns.sd(), ns.maximum());
3284 }
3285 }
3286
3287 void ConcurrentMark::print_summary_info() {
3288 gclog_or_tty->print_cr(" Concurrent marking:");
3289 print_ms_time_info(" ", "init marks", _init_times);
3290 print_ms_time_info(" ", "remarks", _remark_times);
3291 {
3292 print_ms_time_info(" ", "final marks", _remark_mark_times);
3293 print_ms_time_info(" ", "weak refs", _remark_weak_ref_times);
3294
3295 }
3296 print_ms_time_info(" ", "cleanups", _cleanup_times);
3297 gclog_or_tty->print_cr(" Final counting total time = %8.2f s (avg = %8.2f ms).",
3298 _total_counting_time,
3299 (_cleanup_times.num() > 0 ? _total_counting_time * 1000.0 /
3300 (double)_cleanup_times.num()
3301 : 0.0));
3302 if (G1ScrubRemSets) {
3303 gclog_or_tty->print_cr(" RS scrub total time = %8.2f s (avg = %8.2f ms).",
3304 _total_rs_scrub_time,
3305 (_cleanup_times.num() > 0 ? _total_rs_scrub_time * 1000.0 /
3306 (double)_cleanup_times.num()
3307 : 0.0));
3308 }
3309 gclog_or_tty->print_cr(" Total stop_world time = %8.2f s.",
3310 (_init_times.sum() + _remark_times.sum() +
3311 _cleanup_times.sum())/1000.0);
3312 gclog_or_tty->print_cr(" Total concurrent time = %8.2f s "
3313 "(%8.2f s marking).",
3314 cmThread()->vtime_accum(),
3315 cmThread()->vtime_mark_accum());
3316 }
3317
3318 void ConcurrentMark::print_worker_threads_on(outputStream* st) const {
3319 if (use_parallel_marking_threads()) {
3320 _parallel_workers->print_worker_threads_on(st);
3321 }
3322 }
3323
3324 void ConcurrentMark::print_on_error(outputStream* st) const {
3325 st->print_cr("Marking Bits (Prev, Next): (CMBitMap*) " PTR_FORMAT ", (CMBitMap*) " PTR_FORMAT,
3326 p2i(_prevMarkBitMap), p2i(_nextMarkBitMap));
3327 _prevMarkBitMap->print_on_error(st, " Prev Bits: ");
3328 _nextMarkBitMap->print_on_error(st, " Next Bits: ");
3329 }
3330
3331 // We take a break if someone is trying to stop the world.
3332 bool ConcurrentMark::do_yield_check(uint worker_id) {
3333 if (SuspendibleThreadSet::should_yield()) {
3334 if (worker_id == 0) {
3335 _g1h->g1_policy()->record_concurrent_pause();
3336 }
3337 SuspendibleThreadSet::yield();
3338 return true;
3339 } else {
3340 return false;
3341 }
3342 }
3343
3344 bool ConcurrentMark::containing_card_is_marked(void* p) {
3345 size_t offset = pointer_delta(p, _g1h->reserved_region().start(), 1);
3346 return _card_bm.at(offset >> CardTableModRefBS::card_shift);
3347 }
3348
3349 bool ConcurrentMark::containing_cards_are_marked(void* start,
3350 void* last) {
3351 return containing_card_is_marked(start) &&
3352 containing_card_is_marked(last);
3353 }
3354
3355 #ifndef PRODUCT
3356 // for debugging purposes
3357 void ConcurrentMark::print_finger() {
3358 gclog_or_tty->print_cr("heap ["PTR_FORMAT", "PTR_FORMAT"), global finger = "PTR_FORMAT,
3359 p2i(_heap_start), p2i(_heap_end), p2i(_finger));
3360 for (uint i = 0; i < _max_worker_id; ++i) {
3361 gclog_or_tty->print(" %u: " PTR_FORMAT, i, p2i(_tasks[i]->finger()));
3362 }
3363 gclog_or_tty->cr();
3364 }
3365 #endif
3366
3367 void CMTask::scan_object(oop obj) {
3368 assert(_nextMarkBitMap->isMarked((HeapWord*) obj), "invariant");
3369
3370 if (_cm->verbose_high()) {
3371 gclog_or_tty->print_cr("[%u] we're scanning object "PTR_FORMAT,
3372 _worker_id, p2i((void*) obj));
3373 }
3374
3375 size_t obj_size = obj->size();
3376 _words_scanned += obj_size;
3377
3378 obj->oop_iterate(_cm_oop_closure);
3379 statsOnly( ++_objs_scanned );
3380 check_limits();
3381 }
3382
3383 // Closure for iteration over bitmaps
3384 class CMBitMapClosure : public BitMapClosure {
3385 private:
3386 // the bitmap that is being iterated over
3387 CMBitMap* _nextMarkBitMap;
3388 ConcurrentMark* _cm;
3389 CMTask* _task;
3390
3391 public:
3392 CMBitMapClosure(CMTask *task, ConcurrentMark* cm, CMBitMap* nextMarkBitMap) :
3393 _task(task), _cm(cm), _nextMarkBitMap(nextMarkBitMap) { }
3394
3395 bool do_bit(size_t offset) {
3396 HeapWord* addr = _nextMarkBitMap->offsetToHeapWord(offset);
3397 assert(_nextMarkBitMap->isMarked(addr), "invariant");
3398 assert( addr < _cm->finger(), "invariant");
3399
3400 statsOnly( _task->increase_objs_found_on_bitmap() );
3401 assert(addr >= _task->finger(), "invariant");
3402
3403 // We move that task's local finger along.
3404 _task->move_finger_to(addr);
3405
3406 _task->scan_object(oop(addr));
3407 // we only partially drain the local queue and global stack
3408 _task->drain_local_queue(true);
3409 _task->drain_global_stack(true);
3410
3411 // if the has_aborted flag has been raised, we need to bail out of
3412 // the iteration
3413 return !_task->has_aborted();
3414 }
3415 };
3416
3417 // Closure for iterating over objects, currently only used for
3418 // processing SATB buffers.
3419 class CMObjectClosure : public ObjectClosure {
3420 private:
3421 CMTask* _task;
3422
3423 public:
3424 void do_object(oop obj) {
3425 _task->deal_with_reference(obj);
3426 }
3427
3428 CMObjectClosure(CMTask* task) : _task(task) { }
3429 };
3430
3431 G1CMOopClosure::G1CMOopClosure(G1CollectedHeap* g1h,
3432 ConcurrentMark* cm,
3433 CMTask* task)
3434 : _g1h(g1h), _cm(cm), _task(task) {
3435 assert(_ref_processor == NULL, "should be initialized to NULL");
3436
3437 if (G1UseConcMarkReferenceProcessing) {
3438 _ref_processor = g1h->ref_processor_cm();
3439 assert(_ref_processor != NULL, "should not be NULL");
3440 }
3441 }
3442
3443 void CMTask::setup_for_region(HeapRegion* hr) {
3444 assert(hr != NULL,
3445 "claim_region() should have filtered out NULL regions");
3446 assert(!hr->continuesHumongous(),
3447 "claim_region() should have filtered out continues humongous regions");
3448
3449 if (_cm->verbose_low()) {
3450 gclog_or_tty->print_cr("[%u] setting up for region "PTR_FORMAT,
3451 _worker_id, p2i(hr));
3452 }
3453
3454 _curr_region = hr;
3455 _finger = hr->bottom();
3456 update_region_limit();
3457 }
3458
3459 void CMTask::update_region_limit() {
3460 HeapRegion* hr = _curr_region;
3461 HeapWord* bottom = hr->bottom();
3462 HeapWord* limit = hr->next_top_at_mark_start();
3463
3464 if (limit == bottom) {
3465 if (_cm->verbose_low()) {
3466 gclog_or_tty->print_cr("[%u] found an empty region "
3467 "["PTR_FORMAT", "PTR_FORMAT")",
3468 _worker_id, p2i(bottom), p2i(limit));
3469 }
3470 // The region was collected underneath our feet.
3471 // We set the finger to bottom to ensure that the bitmap
3472 // iteration that will follow this will not do anything.
3473 // (this is not a condition that holds when we set the region up,
3474 // as the region is not supposed to be empty in the first place)
3475 _finger = bottom;
3476 } else if (limit >= _region_limit) {
3477 assert(limit >= _finger, "peace of mind");
3478 } else {
3479 assert(limit < _region_limit, "only way to get here");
3480 // This can happen under some pretty unusual circumstances. An
3481 // evacuation pause empties the region underneath our feet (NTAMS
3482 // at bottom). We then do some allocation in the region (NTAMS
3483 // stays at bottom), followed by the region being used as a GC
3484 // alloc region (NTAMS will move to top() and the objects
3485 // originally below it will be grayed). All objects now marked in
3486 // the region are explicitly grayed, if below the global finger,
3487 // and we do not need in fact to scan anything else. So, we simply
3488 // set _finger to be limit to ensure that the bitmap iteration
3489 // doesn't do anything.
3490 _finger = limit;
3491 }
3492
3493 _region_limit = limit;
3494 }
3495
3496 void CMTask::giveup_current_region() {
3497 assert(_curr_region != NULL, "invariant");
3498 if (_cm->verbose_low()) {
3499 gclog_or_tty->print_cr("[%u] giving up region "PTR_FORMAT,
3500 _worker_id, p2i(_curr_region));
3501 }
3502 clear_region_fields();
3503 }
3504
3505 void CMTask::clear_region_fields() {
3506 // Values for these three fields that indicate that we're not
3507 // holding on to a region.
3508 _curr_region = NULL;
3509 _finger = NULL;
3510 _region_limit = NULL;
3511 }
3512
3513 void CMTask::set_cm_oop_closure(G1CMOopClosure* cm_oop_closure) {
3514 if (cm_oop_closure == NULL) {
3515 assert(_cm_oop_closure != NULL, "invariant");
3516 } else {
3517 assert(_cm_oop_closure == NULL, "invariant");
3518 }
3519 _cm_oop_closure = cm_oop_closure;
3520 }
3521
3522 void CMTask::reset(CMBitMap* nextMarkBitMap) {
3523 guarantee(nextMarkBitMap != NULL, "invariant");
3524
3525 if (_cm->verbose_low()) {
3526 gclog_or_tty->print_cr("[%u] resetting", _worker_id);
3527 }
3528
3529 _nextMarkBitMap = nextMarkBitMap;
3530 clear_region_fields();
3531
3532 _calls = 0;
3533 _elapsed_time_ms = 0.0;
3534 _termination_time_ms = 0.0;
3535 _termination_start_time_ms = 0.0;
3536
3537 #if _MARKING_STATS_
3538 _local_pushes = 0;
3539 _local_pops = 0;
3540 _local_max_size = 0;
3541 _objs_scanned = 0;
3542 _global_pushes = 0;
3543 _global_pops = 0;
3544 _global_max_size = 0;
3545 _global_transfers_to = 0;
3546 _global_transfers_from = 0;
3547 _regions_claimed = 0;
3548 _objs_found_on_bitmap = 0;
3549 _satb_buffers_processed = 0;
3550 _steal_attempts = 0;
3551 _steals = 0;
3552 _aborted = 0;
3553 _aborted_overflow = 0;
3554 _aborted_cm_aborted = 0;
3555 _aborted_yield = 0;
3556 _aborted_timed_out = 0;
3557 _aborted_satb = 0;
3558 _aborted_termination = 0;
3559 #endif // _MARKING_STATS_
3560 }
3561
3562 bool CMTask::should_exit_termination() {
3563 regular_clock_call();
3564 // This is called when we are in the termination protocol. We should
3565 // quit if, for some reason, this task wants to abort or the global
3566 // stack is not empty (this means that we can get work from it).
3567 return !_cm->mark_stack_empty() || has_aborted();
3568 }
3569
3570 void CMTask::reached_limit() {
3571 assert(_words_scanned >= _words_scanned_limit ||
3572 _refs_reached >= _refs_reached_limit ,
3573 "shouldn't have been called otherwise");
3574 regular_clock_call();
3575 }
3576
3577 void CMTask::regular_clock_call() {
3578 if (has_aborted()) return;
3579
3580 // First, we need to recalculate the words scanned and refs reached
3581 // limits for the next clock call.
3582 recalculate_limits();
3583
3584 // During the regular clock call we do the following
3585
3586 // (1) If an overflow has been flagged, then we abort.
3587 if (_cm->has_overflown()) {
3588 set_has_aborted();
3589 return;
3590 }
3591
3592 // If we are not concurrent (i.e. we're doing remark) we don't need
3593 // to check anything else. The other steps are only needed during
3594 // the concurrent marking phase.
3595 if (!concurrent()) return;
3596
3597 // (2) If marking has been aborted for Full GC, then we also abort.
3598 if (_cm->has_aborted()) {
3599 set_has_aborted();
3600 statsOnly( ++_aborted_cm_aborted );
3601 return;
3602 }
3603
3604 double curr_time_ms = os::elapsedVTime() * 1000.0;
3605
3606 // (3) If marking stats are enabled, then we update the step history.
3607 #if _MARKING_STATS_
3608 if (_words_scanned >= _words_scanned_limit) {
3609 ++_clock_due_to_scanning;
3610 }
3611 if (_refs_reached >= _refs_reached_limit) {
3612 ++_clock_due_to_marking;
3613 }
3614
3615 double last_interval_ms = curr_time_ms - _interval_start_time_ms;
3616 _interval_start_time_ms = curr_time_ms;
3617 _all_clock_intervals_ms.add(last_interval_ms);
3618
3619 if (_cm->verbose_medium()) {
3620 gclog_or_tty->print_cr("[%u] regular clock, interval = %1.2lfms, "
3621 "scanned = %d%s, refs reached = %d%s",
3622 _worker_id, last_interval_ms,
3623 _words_scanned,
3624 (_words_scanned >= _words_scanned_limit) ? " (*)" : "",
3625 _refs_reached,
3626 (_refs_reached >= _refs_reached_limit) ? " (*)" : "");
3627 }
3628 #endif // _MARKING_STATS_
3629
3630 // (4) We check whether we should yield. If we have to, then we abort.
3631 if (SuspendibleThreadSet::should_yield()) {
3632 // We should yield. To do this we abort the task. The caller is
3633 // responsible for yielding.
3634 set_has_aborted();
3635 statsOnly( ++_aborted_yield );
3636 return;
3637 }
3638
3639 // (5) We check whether we've reached our time quota. If we have,
3640 // then we abort.
3641 double elapsed_time_ms = curr_time_ms - _start_time_ms;
3642 if (elapsed_time_ms > _time_target_ms) {
3643 set_has_aborted();
3644 _has_timed_out = true;
3645 statsOnly( ++_aborted_timed_out );
3646 return;
3647 }
3648
3649 // (6) Finally, we check whether there are enough completed STAB
3650 // buffers available for processing. If there are, we abort.
3651 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3652 if (!_draining_satb_buffers && satb_mq_set.process_completed_buffers()) {
3653 if (_cm->verbose_low()) {
3654 gclog_or_tty->print_cr("[%u] aborting to deal with pending SATB buffers",
3655 _worker_id);
3656 }
3657 // we do need to process SATB buffers, we'll abort and restart
3658 // the marking task to do so
3659 set_has_aborted();
3660 statsOnly( ++_aborted_satb );
3661 return;
3662 }
3663 }
3664
3665 void CMTask::recalculate_limits() {
3666 _real_words_scanned_limit = _words_scanned + words_scanned_period;
3667 _words_scanned_limit = _real_words_scanned_limit;
3668
3669 _real_refs_reached_limit = _refs_reached + refs_reached_period;
3670 _refs_reached_limit = _real_refs_reached_limit;
3671 }
3672
3673 void CMTask::decrease_limits() {
3674 // This is called when we believe that we're going to do an infrequent
3675 // operation which will increase the per byte scanned cost (i.e. move
3676 // entries to/from the global stack). It basically tries to decrease the
3677 // scanning limit so that the clock is called earlier.
3678
3679 if (_cm->verbose_medium()) {
3680 gclog_or_tty->print_cr("[%u] decreasing limits", _worker_id);
3681 }
3682
3683 _words_scanned_limit = _real_words_scanned_limit -
3684 3 * words_scanned_period / 4;
3685 _refs_reached_limit = _real_refs_reached_limit -
3686 3 * refs_reached_period / 4;
3687 }
3688
3689 void CMTask::move_entries_to_global_stack() {
3690 // local array where we'll store the entries that will be popped
3691 // from the local queue
3692 oop buffer[global_stack_transfer_size];
3693
3694 int n = 0;
3695 oop obj;
3696 while (n < global_stack_transfer_size && _task_queue->pop_local(obj)) {
3697 buffer[n] = obj;
3698 ++n;
3699 }
3700
3701 if (n > 0) {
3702 // we popped at least one entry from the local queue
3703
3704 statsOnly( ++_global_transfers_to; _local_pops += n );
3705
3706 if (!_cm->mark_stack_push(buffer, n)) {
3707 if (_cm->verbose_low()) {
3708 gclog_or_tty->print_cr("[%u] aborting due to global stack overflow",
3709 _worker_id);
3710 }
3711 set_has_aborted();
3712 } else {
3713 // the transfer was successful
3714
3715 if (_cm->verbose_medium()) {
3716 gclog_or_tty->print_cr("[%u] pushed %d entries to the global stack",
3717 _worker_id, n);
3718 }
3719 statsOnly( int tmp_size = _cm->mark_stack_size();
3720 if (tmp_size > _global_max_size) {
3721 _global_max_size = tmp_size;
3722 }
3723 _global_pushes += n );
3724 }
3725 }
3726
3727 // this operation was quite expensive, so decrease the limits
3728 decrease_limits();
3729 }
3730
3731 void CMTask::get_entries_from_global_stack() {
3732 // local array where we'll store the entries that will be popped
3733 // from the global stack.
3734 oop buffer[global_stack_transfer_size];
3735 int n;
3736 _cm->mark_stack_pop(buffer, global_stack_transfer_size, &n);
3737 assert(n <= global_stack_transfer_size,
3738 "we should not pop more than the given limit");
3739 if (n > 0) {
3740 // yes, we did actually pop at least one entry
3741
3742 statsOnly( ++_global_transfers_from; _global_pops += n );
3743 if (_cm->verbose_medium()) {
3744 gclog_or_tty->print_cr("[%u] popped %d entries from the global stack",
3745 _worker_id, n);
3746 }
3747 for (int i = 0; i < n; ++i) {
3748 bool success = _task_queue->push(buffer[i]);
3749 // We only call this when the local queue is empty or under a
3750 // given target limit. So, we do not expect this push to fail.
3751 assert(success, "invariant");
3752 }
3753
3754 statsOnly( int tmp_size = _task_queue->size();
3755 if (tmp_size > _local_max_size) {
3756 _local_max_size = tmp_size;
3757 }
3758 _local_pushes += n );
3759 }
3760
3761 // this operation was quite expensive, so decrease the limits
3762 decrease_limits();
3763 }
3764
3765 void CMTask::drain_local_queue(bool partially) {
3766 if (has_aborted()) return;
3767
3768 // Decide what the target size is, depending whether we're going to
3769 // drain it partially (so that other tasks can steal if they run out
3770 // of things to do) or totally (at the very end).
3771 size_t target_size;
3772 if (partially) {
3773 target_size = MIN2((size_t)_task_queue->max_elems()/3, GCDrainStackTargetSize);
3774 } else {
3775 target_size = 0;
3776 }
3777
3778 if (_task_queue->size() > target_size) {
3779 if (_cm->verbose_high()) {
3780 gclog_or_tty->print_cr("[%u] draining local queue, target size = " SIZE_FORMAT,
3781 _worker_id, target_size);
3782 }
3783
3784 oop obj;
3785 bool ret = _task_queue->pop_local(obj);
3786 while (ret) {
3787 statsOnly( ++_local_pops );
3788
3789 if (_cm->verbose_high()) {
3790 gclog_or_tty->print_cr("[%u] popped "PTR_FORMAT, _worker_id,
3791 p2i((void*) obj));
3792 }
3793
3794 assert(_g1h->is_in_g1_reserved((HeapWord*) obj), "invariant" );
3795 assert(!_g1h->is_on_master_free_list(
3796 _g1h->heap_region_containing((HeapWord*) obj)), "invariant");
3797
3798 scan_object(obj);
3799
3800 if (_task_queue->size() <= target_size || has_aborted()) {
3801 ret = false;
3802 } else {
3803 ret = _task_queue->pop_local(obj);
3804 }
3805 }
3806
3807 if (_cm->verbose_high()) {
3808 gclog_or_tty->print_cr("[%u] drained local queue, size = %u",
3809 _worker_id, _task_queue->size());
3810 }
3811 }
3812 }
3813
3814 void CMTask::drain_global_stack(bool partially) {
3815 if (has_aborted()) return;
3816
3817 // We have a policy to drain the local queue before we attempt to
3818 // drain the global stack.
3819 assert(partially || _task_queue->size() == 0, "invariant");
3820
3821 // Decide what the target size is, depending whether we're going to
3822 // drain it partially (so that other tasks can steal if they run out
3823 // of things to do) or totally (at the very end). Notice that,
3824 // because we move entries from the global stack in chunks or
3825 // because another task might be doing the same, we might in fact
3826 // drop below the target. But, this is not a problem.
3827 size_t target_size;
3828 if (partially) {
3829 target_size = _cm->partial_mark_stack_size_target();
3830 } else {
3831 target_size = 0;
3832 }
3833
3834 if (_cm->mark_stack_size() > target_size) {
3835 if (_cm->verbose_low()) {
3836 gclog_or_tty->print_cr("[%u] draining global_stack, target size " SIZE_FORMAT,
3837 _worker_id, target_size);
3838 }
3839
3840 while (!has_aborted() && _cm->mark_stack_size() > target_size) {
3841 get_entries_from_global_stack();
3842 drain_local_queue(partially);
3843 }
3844
3845 if (_cm->verbose_low()) {
3846 gclog_or_tty->print_cr("[%u] drained global stack, size = " SIZE_FORMAT,
3847 _worker_id, _cm->mark_stack_size());
3848 }
3849 }
3850 }
3851
3852 // SATB Queue has several assumptions on whether to call the par or
3853 // non-par versions of the methods. this is why some of the code is
3854 // replicated. We should really get rid of the single-threaded version
3855 // of the code to simplify things.
3856 void CMTask::drain_satb_buffers() {
3857 if (has_aborted()) return;
3858
3859 // We set this so that the regular clock knows that we're in the
3860 // middle of draining buffers and doesn't set the abort flag when it
3861 // notices that SATB buffers are available for draining. It'd be
3862 // very counter productive if it did that. :-)
3863 _draining_satb_buffers = true;
3864
3865 CMObjectClosure oc(this);
3866 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3867 if (G1CollectedHeap::use_parallel_gc_threads()) {
3868 satb_mq_set.set_par_closure(_worker_id, &oc);
3869 } else {
3870 satb_mq_set.set_closure(&oc);
3871 }
3872
3873 // This keeps claiming and applying the closure to completed buffers
3874 // until we run out of buffers or we need to abort.
3875 if (G1CollectedHeap::use_parallel_gc_threads()) {
3876 while (!has_aborted() &&
3877 satb_mq_set.par_apply_closure_to_completed_buffer(_worker_id)) {
3878 if (_cm->verbose_medium()) {
3879 gclog_or_tty->print_cr("[%u] processed an SATB buffer", _worker_id);
3880 }
3881 statsOnly( ++_satb_buffers_processed );
3882 regular_clock_call();
3883 }
3884 } else {
3885 while (!has_aborted() &&
3886 satb_mq_set.apply_closure_to_completed_buffer()) {
3887 if (_cm->verbose_medium()) {
3888 gclog_or_tty->print_cr("[%u] processed an SATB buffer", _worker_id);
3889 }
3890 statsOnly( ++_satb_buffers_processed );
3891 regular_clock_call();
3892 }
3893 }
3894
3895 if (!concurrent() && !has_aborted()) {
3896 // We should only do this during remark.
3897 if (G1CollectedHeap::use_parallel_gc_threads()) {
3898 satb_mq_set.par_iterate_closure_all_threads(_worker_id);
3899 } else {
3900 satb_mq_set.iterate_closure_all_threads();
3901 }
3902 }
3903
3904 _draining_satb_buffers = false;
3905
3906 assert(has_aborted() ||
3907 concurrent() ||
3908 satb_mq_set.completed_buffers_num() == 0, "invariant");
3909
3910 if (G1CollectedHeap::use_parallel_gc_threads()) {
3911 satb_mq_set.set_par_closure(_worker_id, NULL);
3912 } else {
3913 satb_mq_set.set_closure(NULL);
3914 }
3915
3916 // again, this was a potentially expensive operation, decrease the
3917 // limits to get the regular clock call early
3918 decrease_limits();
3919 }
3920
3921 void CMTask::print_stats() {
3922 gclog_or_tty->print_cr("Marking Stats, task = %u, calls = %d",
3923 _worker_id, _calls);
3924 gclog_or_tty->print_cr(" Elapsed time = %1.2lfms, Termination time = %1.2lfms",
3925 _elapsed_time_ms, _termination_time_ms);
3926 gclog_or_tty->print_cr(" Step Times (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms",
3927 _step_times_ms.num(), _step_times_ms.avg(),
3928 _step_times_ms.sd());
3929 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms",
3930 _step_times_ms.maximum(), _step_times_ms.sum());
3931
3932 #if _MARKING_STATS_
3933 gclog_or_tty->print_cr(" Clock Intervals (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms",
3934 _all_clock_intervals_ms.num(), _all_clock_intervals_ms.avg(),
3935 _all_clock_intervals_ms.sd());
3936 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms",
3937 _all_clock_intervals_ms.maximum(),
3938 _all_clock_intervals_ms.sum());
3939 gclog_or_tty->print_cr(" Clock Causes (cum): scanning = %d, marking = %d",
3940 _clock_due_to_scanning, _clock_due_to_marking);
3941 gclog_or_tty->print_cr(" Objects: scanned = %d, found on the bitmap = %d",
3942 _objs_scanned, _objs_found_on_bitmap);
3943 gclog_or_tty->print_cr(" Local Queue: pushes = %d, pops = %d, max size = %d",
3944 _local_pushes, _local_pops, _local_max_size);
3945 gclog_or_tty->print_cr(" Global Stack: pushes = %d, pops = %d, max size = %d",
3946 _global_pushes, _global_pops, _global_max_size);
3947 gclog_or_tty->print_cr(" transfers to = %d, transfers from = %d",
3948 _global_transfers_to,_global_transfers_from);
3949 gclog_or_tty->print_cr(" Regions: claimed = %d", _regions_claimed);
3950 gclog_or_tty->print_cr(" SATB buffers: processed = %d", _satb_buffers_processed);
3951 gclog_or_tty->print_cr(" Steals: attempts = %d, successes = %d",
3952 _steal_attempts, _steals);
3953 gclog_or_tty->print_cr(" Aborted: %d, due to", _aborted);
3954 gclog_or_tty->print_cr(" overflow: %d, global abort: %d, yield: %d",
3955 _aborted_overflow, _aborted_cm_aborted, _aborted_yield);
3956 gclog_or_tty->print_cr(" time out: %d, SATB: %d, termination: %d",
3957 _aborted_timed_out, _aborted_satb, _aborted_termination);
3958 #endif // _MARKING_STATS_
3959 }
3960
3961 /*****************************************************************************
3962
3963 The do_marking_step(time_target_ms, ...) method is the building
3964 block of the parallel marking framework. It can be called in parallel
3965 with other invocations of do_marking_step() on different tasks
3966 (but only one per task, obviously) and concurrently with the
3967 mutator threads, or during remark, hence it eliminates the need
3968 for two versions of the code. When called during remark, it will
3969 pick up from where the task left off during the concurrent marking
3970 phase. Interestingly, tasks are also claimable during evacuation
3971 pauses too, since do_marking_step() ensures that it aborts before
3972 it needs to yield.
3973
3974 The data structures that it uses to do marking work are the
3975 following:
3976
3977 (1) Marking Bitmap. If there are gray objects that appear only
3978 on the bitmap (this happens either when dealing with an overflow
3979 or when the initial marking phase has simply marked the roots
3980 and didn't push them on the stack), then tasks claim heap
3981 regions whose bitmap they then scan to find gray objects. A
3982 global finger indicates where the end of the last claimed region
3983 is. A local finger indicates how far into the region a task has
3984 scanned. The two fingers are used to determine how to gray an
3985 object (i.e. whether simply marking it is OK, as it will be
3986 visited by a task in the future, or whether it needs to be also
3987 pushed on a stack).
3988
3989 (2) Local Queue. The local queue of the task which is accessed
3990 reasonably efficiently by the task. Other tasks can steal from
3991 it when they run out of work. Throughout the marking phase, a
3992 task attempts to keep its local queue short but not totally
3993 empty, so that entries are available for stealing by other
3994 tasks. Only when there is no more work, a task will totally
3995 drain its local queue.
3996
3997 (3) Global Mark Stack. This handles local queue overflow. During
3998 marking only sets of entries are moved between it and the local
3999 queues, as access to it requires a mutex and more fine-grain
4000 interaction with it which might cause contention. If it
4001 overflows, then the marking phase should restart and iterate
4002 over the bitmap to identify gray objects. Throughout the marking
4003 phase, tasks attempt to keep the global mark stack at a small
4004 length but not totally empty, so that entries are available for
4005 popping by other tasks. Only when there is no more work, tasks
4006 will totally drain the global mark stack.
4007
4008 (4) SATB Buffer Queue. This is where completed SATB buffers are
4009 made available. Buffers are regularly removed from this queue
4010 and scanned for roots, so that the queue doesn't get too
4011 long. During remark, all completed buffers are processed, as
4012 well as the filled in parts of any uncompleted buffers.
4013
4014 The do_marking_step() method tries to abort when the time target
4015 has been reached. There are a few other cases when the
4016 do_marking_step() method also aborts:
4017
4018 (1) When the marking phase has been aborted (after a Full GC).
4019
4020 (2) When a global overflow (on the global stack) has been
4021 triggered. Before the task aborts, it will actually sync up with
4022 the other tasks to ensure that all the marking data structures
4023 (local queues, stacks, fingers etc.) are re-initialized so that
4024 when do_marking_step() completes, the marking phase can
4025 immediately restart.
4026
4027 (3) When enough completed SATB buffers are available. The
4028 do_marking_step() method only tries to drain SATB buffers right
4029 at the beginning. So, if enough buffers are available, the
4030 marking step aborts and the SATB buffers are processed at
4031 the beginning of the next invocation.
4032
4033 (4) To yield. when we have to yield then we abort and yield
4034 right at the end of do_marking_step(). This saves us from a lot
4035 of hassle as, by yielding we might allow a Full GC. If this
4036 happens then objects will be compacted underneath our feet, the
4037 heap might shrink, etc. We save checking for this by just
4038 aborting and doing the yield right at the end.
4039
4040 From the above it follows that the do_marking_step() method should
4041 be called in a loop (or, otherwise, regularly) until it completes.
4042
4043 If a marking step completes without its has_aborted() flag being
4044 true, it means it has completed the current marking phase (and
4045 also all other marking tasks have done so and have all synced up).
4046
4047 A method called regular_clock_call() is invoked "regularly" (in
4048 sub ms intervals) throughout marking. It is this clock method that
4049 checks all the abort conditions which were mentioned above and
4050 decides when the task should abort. A work-based scheme is used to
4051 trigger this clock method: when the number of object words the
4052 marking phase has scanned or the number of references the marking
4053 phase has visited reach a given limit. Additional invocations to
4054 the method clock have been planted in a few other strategic places
4055 too. The initial reason for the clock method was to avoid calling
4056 vtime too regularly, as it is quite expensive. So, once it was in
4057 place, it was natural to piggy-back all the other conditions on it
4058 too and not constantly check them throughout the code.
4059
4060 If do_termination is true then do_marking_step will enter its
4061 termination protocol.
4062
4063 The value of is_serial must be true when do_marking_step is being
4064 called serially (i.e. by the VMThread) and do_marking_step should
4065 skip any synchronization in the termination and overflow code.
4066 Examples include the serial remark code and the serial reference
4067 processing closures.
4068
4069 The value of is_serial must be false when do_marking_step is
4070 being called by any of the worker threads in a work gang.
4071 Examples include the concurrent marking code (CMMarkingTask),
4072 the MT remark code, and the MT reference processing closures.
4073
4074 *****************************************************************************/
4075
4076 void CMTask::do_marking_step(double time_target_ms,
4077 bool do_termination,
4078 bool is_serial) {
4079 assert(time_target_ms >= 1.0, "minimum granularity is 1ms");
4080 assert(concurrent() == _cm->concurrent(), "they should be the same");
4081
4082 G1CollectorPolicy* g1_policy = _g1h->g1_policy();
4083 assert(_task_queues != NULL, "invariant");
4084 assert(_task_queue != NULL, "invariant");
4085 assert(_task_queues->queue(_worker_id) == _task_queue, "invariant");
4086
4087 assert(!_claimed,
4088 "only one thread should claim this task at any one time");
4089
4090 // OK, this doesn't safeguard again all possible scenarios, as it is
4091 // possible for two threads to set the _claimed flag at the same
4092 // time. But it is only for debugging purposes anyway and it will
4093 // catch most problems.
4094 _claimed = true;
4095
4096 _start_time_ms = os::elapsedVTime() * 1000.0;
4097 statsOnly( _interval_start_time_ms = _start_time_ms );
4098
4099 // If do_stealing is true then do_marking_step will attempt to
4100 // steal work from the other CMTasks. It only makes sense to
4101 // enable stealing when the termination protocol is enabled
4102 // and do_marking_step() is not being called serially.
4103 bool do_stealing = do_termination && !is_serial;
4104
4105 double diff_prediction_ms =
4106 g1_policy->get_new_prediction(&_marking_step_diffs_ms);
4107 _time_target_ms = time_target_ms - diff_prediction_ms;
4108
4109 // set up the variables that are used in the work-based scheme to
4110 // call the regular clock method
4111 _words_scanned = 0;
4112 _refs_reached = 0;
4113 recalculate_limits();
4114
4115 // clear all flags
4116 clear_has_aborted();
4117 _has_timed_out = false;
4118 _draining_satb_buffers = false;
4119
4120 ++_calls;
4121
4122 if (_cm->verbose_low()) {
4123 gclog_or_tty->print_cr("[%u] >>>>>>>>>> START, call = %d, "
4124 "target = %1.2lfms >>>>>>>>>>",
4125 _worker_id, _calls, _time_target_ms);
4126 }
4127
4128 // Set up the bitmap and oop closures. Anything that uses them is
4129 // eventually called from this method, so it is OK to allocate these
4130 // statically.
4131 CMBitMapClosure bitmap_closure(this, _cm, _nextMarkBitMap);
4132 G1CMOopClosure cm_oop_closure(_g1h, _cm, this);
4133 set_cm_oop_closure(&cm_oop_closure);
4134
4135 if (_cm->has_overflown()) {
4136 // This can happen if the mark stack overflows during a GC pause
4137 // and this task, after a yield point, restarts. We have to abort
4138 // as we need to get into the overflow protocol which happens
4139 // right at the end of this task.
4140 set_has_aborted();
4141 }
4142
4143 // First drain any available SATB buffers. After this, we will not
4144 // look at SATB buffers before the next invocation of this method.
4145 // If enough completed SATB buffers are queued up, the regular clock
4146 // will abort this task so that it restarts.
4147 drain_satb_buffers();
4148 // ...then partially drain the local queue and the global stack
4149 drain_local_queue(true);
4150 drain_global_stack(true);
4151
4152 do {
4153 if (!has_aborted() && _curr_region != NULL) {
4154 // This means that we're already holding on to a region.
4155 assert(_finger != NULL, "if region is not NULL, then the finger "
4156 "should not be NULL either");
4157
4158 // We might have restarted this task after an evacuation pause
4159 // which might have evacuated the region we're holding on to
4160 // underneath our feet. Let's read its limit again to make sure
4161 // that we do not iterate over a region of the heap that
4162 // contains garbage (update_region_limit() will also move
4163 // _finger to the start of the region if it is found empty).
4164 update_region_limit();
4165 // We will start from _finger not from the start of the region,
4166 // as we might be restarting this task after aborting half-way
4167 // through scanning this region. In this case, _finger points to
4168 // the address where we last found a marked object. If this is a
4169 // fresh region, _finger points to start().
4170 MemRegion mr = MemRegion(_finger, _region_limit);
4171
4172 if (_cm->verbose_low()) {
4173 gclog_or_tty->print_cr("[%u] we're scanning part "
4174 "["PTR_FORMAT", "PTR_FORMAT") "
4175 "of region "HR_FORMAT,
4176 _worker_id, p2i(_finger), p2i(_region_limit),
4177 HR_FORMAT_PARAMS(_curr_region));
4178 }
4179
4180 assert(!_curr_region->isHumongous() || mr.start() == _curr_region->bottom(),
4181 "humongous regions should go around loop once only");
4182
4183 // Some special cases:
4184 // If the memory region is empty, we can just give up the region.
4185 // If the current region is humongous then we only need to check
4186 // the bitmap for the bit associated with the start of the object,
4187 // scan the object if it's live, and give up the region.
4188 // Otherwise, let's iterate over the bitmap of the part of the region
4189 // that is left.
4190 // If the iteration is successful, give up the region.
4191 if (mr.is_empty()) {
4192 giveup_current_region();
4193 regular_clock_call();
4194 } else if (_curr_region->isHumongous() && mr.start() == _curr_region->bottom()) {
4195 if (_nextMarkBitMap->isMarked(mr.start())) {
4196 // The object is marked - apply the closure
4197 BitMap::idx_t offset = _nextMarkBitMap->heapWordToOffset(mr.start());
4198 bitmap_closure.do_bit(offset);
4199 }
4200 // Even if this task aborted while scanning the humongous object
4201 // we can (and should) give up the current region.
4202 giveup_current_region();
4203 regular_clock_call();
4204 } else if (_nextMarkBitMap->iterate(&bitmap_closure, mr)) {
4205 giveup_current_region();
4206 regular_clock_call();
4207 } else {
4208 assert(has_aborted(), "currently the only way to do so");
4209 // The only way to abort the bitmap iteration is to return
4210 // false from the do_bit() method. However, inside the
4211 // do_bit() method we move the _finger to point to the
4212 // object currently being looked at. So, if we bail out, we
4213 // have definitely set _finger to something non-null.
4214 assert(_finger != NULL, "invariant");
4215
4216 // Region iteration was actually aborted. So now _finger
4217 // points to the address of the object we last scanned. If we
4218 // leave it there, when we restart this task, we will rescan
4219 // the object. It is easy to avoid this. We move the finger by
4220 // enough to point to the next possible object header (the
4221 // bitmap knows by how much we need to move it as it knows its
4222 // granularity).
4223 assert(_finger < _region_limit, "invariant");
4224 HeapWord* new_finger = _nextMarkBitMap->nextObject(_finger);
4225 // Check if bitmap iteration was aborted while scanning the last object
4226 if (new_finger >= _region_limit) {
4227 giveup_current_region();
4228 } else {
4229 move_finger_to(new_finger);
4230 }
4231 }
4232 }
4233 // At this point we have either completed iterating over the
4234 // region we were holding on to, or we have aborted.
4235
4236 // We then partially drain the local queue and the global stack.
4237 // (Do we really need this?)
4238 drain_local_queue(true);
4239 drain_global_stack(true);
4240
4241 // Read the note on the claim_region() method on why it might
4242 // return NULL with potentially more regions available for
4243 // claiming and why we have to check out_of_regions() to determine
4244 // whether we're done or not.
4245 while (!has_aborted() && _curr_region == NULL && !_cm->out_of_regions()) {
4246 // We are going to try to claim a new region. We should have
4247 // given up on the previous one.
4248 // Separated the asserts so that we know which one fires.
4249 assert(_curr_region == NULL, "invariant");
4250 assert(_finger == NULL, "invariant");
4251 assert(_region_limit == NULL, "invariant");
4252 if (_cm->verbose_low()) {
4253 gclog_or_tty->print_cr("[%u] trying to claim a new region", _worker_id);
4254 }
4255 HeapRegion* claimed_region = _cm->claim_region(_worker_id);
4256 if (claimed_region != NULL) {
4257 // Yes, we managed to claim one
4258 statsOnly( ++_regions_claimed );
4259
4260 if (_cm->verbose_low()) {
4261 gclog_or_tty->print_cr("[%u] we successfully claimed "
4262 "region "PTR_FORMAT,
4263 _worker_id, p2i(claimed_region));
4264 }
4265
4266 setup_for_region(claimed_region);
4267 assert(_curr_region == claimed_region, "invariant");
4268 }
4269 // It is important to call the regular clock here. It might take
4270 // a while to claim a region if, for example, we hit a large
4271 // block of empty regions. So we need to call the regular clock
4272 // method once round the loop to make sure it's called
4273 // frequently enough.
4274 regular_clock_call();
4275 }
4276
4277 if (!has_aborted() && _curr_region == NULL) {
4278 assert(_cm->out_of_regions(),
4279 "at this point we should be out of regions");
4280 }
4281 } while ( _curr_region != NULL && !has_aborted());
4282
4283 if (!has_aborted()) {
4284 // We cannot check whether the global stack is empty, since other
4285 // tasks might be pushing objects to it concurrently.
4286 assert(_cm->out_of_regions(),
4287 "at this point we should be out of regions");
4288
4289 if (_cm->verbose_low()) {
4290 gclog_or_tty->print_cr("[%u] all regions claimed", _worker_id);
4291 }
4292
4293 // Try to reduce the number of available SATB buffers so that
4294 // remark has less work to do.
4295 drain_satb_buffers();
4296 }
4297
4298 // Since we've done everything else, we can now totally drain the
4299 // local queue and global stack.
4300 drain_local_queue(false);
4301 drain_global_stack(false);
4302
4303 // Attempt at work stealing from other task's queues.
4304 if (do_stealing && !has_aborted()) {
4305 // We have not aborted. This means that we have finished all that
4306 // we could. Let's try to do some stealing...
4307
4308 // We cannot check whether the global stack is empty, since other
4309 // tasks might be pushing objects to it concurrently.
4310 assert(_cm->out_of_regions() && _task_queue->size() == 0,
4311 "only way to reach here");
4312
4313 if (_cm->verbose_low()) {
4314 gclog_or_tty->print_cr("[%u] starting to steal", _worker_id);
4315 }
4316
4317 while (!has_aborted()) {
4318 oop obj;
4319 statsOnly( ++_steal_attempts );
4320
4321 if (_cm->try_stealing(_worker_id, &_hash_seed, obj)) {
4322 if (_cm->verbose_medium()) {
4323 gclog_or_tty->print_cr("[%u] stolen "PTR_FORMAT" successfully",
4324 _worker_id, p2i((void*) obj));
4325 }
4326
4327 statsOnly( ++_steals );
4328
4329 assert(_nextMarkBitMap->isMarked((HeapWord*) obj),
4330 "any stolen object should be marked");
4331 scan_object(obj);
4332
4333 // And since we're towards the end, let's totally drain the
4334 // local queue and global stack.
4335 drain_local_queue(false);
4336 drain_global_stack(false);
4337 } else {
4338 break;
4339 }
4340 }
4341 }
4342
4343 // If we are about to wrap up and go into termination, check if we
4344 // should raise the overflow flag.
4345 if (do_termination && !has_aborted()) {
4346 if (_cm->force_overflow()->should_force()) {
4347 _cm->set_has_overflown();
4348 regular_clock_call();
4349 }
4350 }
4351
4352 // We still haven't aborted. Now, let's try to get into the
4353 // termination protocol.
4354 if (do_termination && !has_aborted()) {
4355 // We cannot check whether the global stack is empty, since other
4356 // tasks might be concurrently pushing objects on it.
4357 // Separated the asserts so that we know which one fires.
4358 assert(_cm->out_of_regions(), "only way to reach here");
4359 assert(_task_queue->size() == 0, "only way to reach here");
4360
4361 if (_cm->verbose_low()) {
4362 gclog_or_tty->print_cr("[%u] starting termination protocol", _worker_id);
4363 }
4364
4365 _termination_start_time_ms = os::elapsedVTime() * 1000.0;
4366
4367 // The CMTask class also extends the TerminatorTerminator class,
4368 // hence its should_exit_termination() method will also decide
4369 // whether to exit the termination protocol or not.
4370 bool finished = (is_serial ||
4371 _cm->terminator()->offer_termination(this));
4372 double termination_end_time_ms = os::elapsedVTime() * 1000.0;
4373 _termination_time_ms +=
4374 termination_end_time_ms - _termination_start_time_ms;
4375
4376 if (finished) {
4377 // We're all done.
4378
4379 if (_worker_id == 0) {
4380 // let's allow task 0 to do this
4381 if (concurrent()) {
4382 assert(_cm->concurrent_marking_in_progress(), "invariant");
4383 // we need to set this to false before the next
4384 // safepoint. This way we ensure that the marking phase
4385 // doesn't observe any more heap expansions.
4386 _cm->clear_concurrent_marking_in_progress();
4387 }
4388 }
4389
4390 // We can now guarantee that the global stack is empty, since
4391 // all other tasks have finished. We separated the guarantees so
4392 // that, if a condition is false, we can immediately find out
4393 // which one.
4394 guarantee(_cm->out_of_regions(), "only way to reach here");
4395 guarantee(_cm->mark_stack_empty(), "only way to reach here");
4396 guarantee(_task_queue->size() == 0, "only way to reach here");
4397 guarantee(!_cm->has_overflown(), "only way to reach here");
4398 guarantee(!_cm->mark_stack_overflow(), "only way to reach here");
4399
4400 if (_cm->verbose_low()) {
4401 gclog_or_tty->print_cr("[%u] all tasks terminated", _worker_id);
4402 }
4403 } else {
4404 // Apparently there's more work to do. Let's abort this task. It
4405 // will restart it and we can hopefully find more things to do.
4406
4407 if (_cm->verbose_low()) {
4408 gclog_or_tty->print_cr("[%u] apparently there is more work to do",
4409 _worker_id);
4410 }
4411
4412 set_has_aborted();
4413 statsOnly( ++_aborted_termination );
4414 }
4415 }
4416
4417 // Mainly for debugging purposes to make sure that a pointer to the
4418 // closure which was statically allocated in this frame doesn't
4419 // escape it by accident.
4420 set_cm_oop_closure(NULL);
4421 double end_time_ms = os::elapsedVTime() * 1000.0;
4422 double elapsed_time_ms = end_time_ms - _start_time_ms;
4423 // Update the step history.
4424 _step_times_ms.add(elapsed_time_ms);
4425
4426 if (has_aborted()) {
4427 // The task was aborted for some reason.
4428
4429 statsOnly( ++_aborted );
4430
4431 if (_has_timed_out) {
4432 double diff_ms = elapsed_time_ms - _time_target_ms;
4433 // Keep statistics of how well we did with respect to hitting
4434 // our target only if we actually timed out (if we aborted for
4435 // other reasons, then the results might get skewed).
4436 _marking_step_diffs_ms.add(diff_ms);
4437 }
4438
4439 if (_cm->has_overflown()) {
4440 // This is the interesting one. We aborted because a global
4441 // overflow was raised. This means we have to restart the
4442 // marking phase and start iterating over regions. However, in
4443 // order to do this we have to make sure that all tasks stop
4444 // what they are doing and re-initialize in a safe manner. We
4445 // will achieve this with the use of two barrier sync points.
4446
4447 if (_cm->verbose_low()) {
4448 gclog_or_tty->print_cr("[%u] detected overflow", _worker_id);
4449 }
4450
4451 if (!is_serial) {
4452 // We only need to enter the sync barrier if being called
4453 // from a parallel context
4454 _cm->enter_first_sync_barrier(_worker_id);
4455
4456 // When we exit this sync barrier we know that all tasks have
4457 // stopped doing marking work. So, it's now safe to
4458 // re-initialize our data structures. At the end of this method,
4459 // task 0 will clear the global data structures.
4460 }
4461
4462 statsOnly( ++_aborted_overflow );
4463
4464 // We clear the local state of this task...
4465 clear_region_fields();
4466
4467 if (!is_serial) {
4468 // ...and enter the second barrier.
4469 _cm->enter_second_sync_barrier(_worker_id);
4470 }
4471 // At this point, if we're during the concurrent phase of
4472 // marking, everything has been re-initialized and we're
4473 // ready to restart.
4474 }
4475
4476 if (_cm->verbose_low()) {
4477 gclog_or_tty->print_cr("[%u] <<<<<<<<<< ABORTING, target = %1.2lfms, "
4478 "elapsed = %1.2lfms <<<<<<<<<<",
4479 _worker_id, _time_target_ms, elapsed_time_ms);
4480 if (_cm->has_aborted()) {
4481 gclog_or_tty->print_cr("[%u] ========== MARKING ABORTED ==========",
4482 _worker_id);
4483 }
4484 }
4485 } else {
4486 if (_cm->verbose_low()) {
4487 gclog_or_tty->print_cr("[%u] <<<<<<<<<< FINISHED, target = %1.2lfms, "
4488 "elapsed = %1.2lfms <<<<<<<<<<",
4489 _worker_id, _time_target_ms, elapsed_time_ms);
4490 }
4491 }
4492
4493 _claimed = false;
4494 }
4495
4496 CMTask::CMTask(uint worker_id,
4497 ConcurrentMark* cm,
4498 size_t* marked_bytes,
4499 BitMap* card_bm,
4500 CMTaskQueue* task_queue,
4501 CMTaskQueueSet* task_queues)
4502 : _g1h(G1CollectedHeap::heap()),
4503 _worker_id(worker_id), _cm(cm),
4504 _claimed(false),
4505 _nextMarkBitMap(NULL), _hash_seed(17),
4506 _task_queue(task_queue),
4507 _task_queues(task_queues),
4508 _cm_oop_closure(NULL),
4509 _marked_bytes_array(marked_bytes),
4510 _card_bm(card_bm) {
4511 guarantee(task_queue != NULL, "invariant");
4512 guarantee(task_queues != NULL, "invariant");
4513
4514 statsOnly( _clock_due_to_scanning = 0;
4515 _clock_due_to_marking = 0 );
4516
4517 _marking_step_diffs_ms.add(0.5);
4518 }
4519
4520 // These are formatting macros that are used below to ensure
4521 // consistent formatting. The *_H_* versions are used to format the
4522 // header for a particular value and they should be kept consistent
4523 // with the corresponding macro. Also note that most of the macros add
4524 // the necessary white space (as a prefix) which makes them a bit
4525 // easier to compose.
4526
4527 // All the output lines are prefixed with this string to be able to
4528 // identify them easily in a large log file.
4529 #define G1PPRL_LINE_PREFIX "###"
4530
4531 #define G1PPRL_ADDR_BASE_FORMAT " "PTR_FORMAT"-"PTR_FORMAT
4532 #ifdef _LP64
4533 #define G1PPRL_ADDR_BASE_H_FORMAT " %37s"
4534 #else // _LP64
4535 #define G1PPRL_ADDR_BASE_H_FORMAT " %21s"
4536 #endif // _LP64
4537
4538 // For per-region info
4539 #define G1PPRL_TYPE_FORMAT " %-4s"
4540 #define G1PPRL_TYPE_H_FORMAT " %4s"
4541 #define G1PPRL_BYTE_FORMAT " "SIZE_FORMAT_W(9)
4542 #define G1PPRL_BYTE_H_FORMAT " %9s"
4543 #define G1PPRL_DOUBLE_FORMAT " %14.1f"
4544 #define G1PPRL_DOUBLE_H_FORMAT " %14s"
4545
4546 // For summary info
4547 #define G1PPRL_SUM_ADDR_FORMAT(tag) " "tag":"G1PPRL_ADDR_BASE_FORMAT
4548 #define G1PPRL_SUM_BYTE_FORMAT(tag) " "tag": "SIZE_FORMAT
4549 #define G1PPRL_SUM_MB_FORMAT(tag) " "tag": %1.2f MB"
4550 #define G1PPRL_SUM_MB_PERC_FORMAT(tag) G1PPRL_SUM_MB_FORMAT(tag)" / %1.2f %%"
4551
4552 G1PrintRegionLivenessInfoClosure::
4553 G1PrintRegionLivenessInfoClosure(outputStream* out, const char* phase_name)
4554 : _out(out),
4555 _total_used_bytes(0), _total_capacity_bytes(0),
4556 _total_prev_live_bytes(0), _total_next_live_bytes(0),
4557 _hum_used_bytes(0), _hum_capacity_bytes(0),
4558 _hum_prev_live_bytes(0), _hum_next_live_bytes(0),
4559 _total_remset_bytes(0), _total_strong_code_roots_bytes(0) {
4560 G1CollectedHeap* g1h = G1CollectedHeap::heap();
4561 MemRegion g1_committed = g1h->g1_committed();
4562 MemRegion g1_reserved = g1h->g1_reserved();
4563 double now = os::elapsedTime();
4564
4565 // Print the header of the output.
4566 _out->cr();
4567 _out->print_cr(G1PPRL_LINE_PREFIX" PHASE %s @ %1.3f", phase_name, now);
4568 _out->print_cr(G1PPRL_LINE_PREFIX" HEAP"
4569 G1PPRL_SUM_ADDR_FORMAT("committed")
4570 G1PPRL_SUM_ADDR_FORMAT("reserved")
4571 G1PPRL_SUM_BYTE_FORMAT("region-size"),
4572 p2i(g1_committed.start()), p2i(g1_committed.end()),
4573 p2i(g1_reserved.start()), p2i(g1_reserved.end()),
4574 HeapRegion::GrainBytes);
4575 _out->print_cr(G1PPRL_LINE_PREFIX);
4576 _out->print_cr(G1PPRL_LINE_PREFIX
4577 G1PPRL_TYPE_H_FORMAT
4578 G1PPRL_ADDR_BASE_H_FORMAT
4579 G1PPRL_BYTE_H_FORMAT
4580 G1PPRL_BYTE_H_FORMAT
4581 G1PPRL_BYTE_H_FORMAT
4582 G1PPRL_DOUBLE_H_FORMAT
4583 G1PPRL_BYTE_H_FORMAT
4584 G1PPRL_BYTE_H_FORMAT,
4585 "type", "address-range",
4586 "used", "prev-live", "next-live", "gc-eff",
4587 "remset", "code-roots");
4588 _out->print_cr(G1PPRL_LINE_PREFIX
4589 G1PPRL_TYPE_H_FORMAT
4590 G1PPRL_ADDR_BASE_H_FORMAT
4591 G1PPRL_BYTE_H_FORMAT
4592 G1PPRL_BYTE_H_FORMAT
4593 G1PPRL_BYTE_H_FORMAT
4594 G1PPRL_DOUBLE_H_FORMAT
4595 G1PPRL_BYTE_H_FORMAT
4596 G1PPRL_BYTE_H_FORMAT,
4597 "", "",
4598 "(bytes)", "(bytes)", "(bytes)", "(bytes/ms)",
4599 "(bytes)", "(bytes)");
4600 }
4601
4602 // It takes as a parameter a reference to one of the _hum_* fields, it
4603 // deduces the corresponding value for a region in a humongous region
4604 // series (either the region size, or what's left if the _hum_* field
4605 // is < the region size), and updates the _hum_* field accordingly.
4606 size_t G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* hum_bytes) {
4607 size_t bytes = 0;
4608 // The > 0 check is to deal with the prev and next live bytes which
4609 // could be 0.
4610 if (*hum_bytes > 0) {
4611 bytes = MIN2(HeapRegion::GrainBytes, *hum_bytes);
4612 *hum_bytes -= bytes;
4613 }
4614 return bytes;
4615 }
4616
4617 // It deduces the values for a region in a humongous region series
4618 // from the _hum_* fields and updates those accordingly. It assumes
4619 // that that _hum_* fields have already been set up from the "starts
4620 // humongous" region and we visit the regions in address order.
4621 void G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* used_bytes,
4622 size_t* capacity_bytes,
4623 size_t* prev_live_bytes,
4624 size_t* next_live_bytes) {
4625 assert(_hum_used_bytes > 0 && _hum_capacity_bytes > 0, "pre-condition");
4626 *used_bytes = get_hum_bytes(&_hum_used_bytes);
4627 *capacity_bytes = get_hum_bytes(&_hum_capacity_bytes);
4628 *prev_live_bytes = get_hum_bytes(&_hum_prev_live_bytes);
4629 *next_live_bytes = get_hum_bytes(&_hum_next_live_bytes);
4630 }
4631
4632 bool G1PrintRegionLivenessInfoClosure::doHeapRegion(HeapRegion* r) {
4633 const char* type = "";
4634 HeapWord* bottom = r->bottom();
4635 HeapWord* end = r->end();
4636 size_t capacity_bytes = r->capacity();
4637 size_t used_bytes = r->used();
4638 size_t prev_live_bytes = r->live_bytes();
4639 size_t next_live_bytes = r->next_live_bytes();
4640 double gc_eff = r->gc_efficiency();
4641 size_t remset_bytes = r->rem_set()->mem_size();
4642 size_t strong_code_roots_bytes = r->rem_set()->strong_code_roots_mem_size();
4643
4644 if (r->used() == 0) {
4645 type = "FREE";
4646 } else if (r->is_survivor()) {
4647 type = "SURV";
4648 } else if (r->is_young()) {
4649 type = "EDEN";
4650 } else if (r->startsHumongous()) {
4651 type = "HUMS";
4652
4653 assert(_hum_used_bytes == 0 && _hum_capacity_bytes == 0 &&
4654 _hum_prev_live_bytes == 0 && _hum_next_live_bytes == 0,
4655 "they should have been zeroed after the last time we used them");
4656 // Set up the _hum_* fields.
4657 _hum_capacity_bytes = capacity_bytes;
4658 _hum_used_bytes = used_bytes;
4659 _hum_prev_live_bytes = prev_live_bytes;
4660 _hum_next_live_bytes = next_live_bytes;
4661 get_hum_bytes(&used_bytes, &capacity_bytes,
4662 &prev_live_bytes, &next_live_bytes);
4663 end = bottom + HeapRegion::GrainWords;
4664 } else if (r->continuesHumongous()) {
4665 type = "HUMC";
4666 get_hum_bytes(&used_bytes, &capacity_bytes,
4667 &prev_live_bytes, &next_live_bytes);
4668 assert(end == bottom + HeapRegion::GrainWords, "invariant");
4669 } else {
4670 type = "OLD";
4671 }
4672
4673 _total_used_bytes += used_bytes;
4674 _total_capacity_bytes += capacity_bytes;
4675 _total_prev_live_bytes += prev_live_bytes;
4676 _total_next_live_bytes += next_live_bytes;
4677 _total_remset_bytes += remset_bytes;
4678 _total_strong_code_roots_bytes += strong_code_roots_bytes;
4679
4680 // Print a line for this particular region.
4681 _out->print_cr(G1PPRL_LINE_PREFIX
4682 G1PPRL_TYPE_FORMAT
4683 G1PPRL_ADDR_BASE_FORMAT
4684 G1PPRL_BYTE_FORMAT
4685 G1PPRL_BYTE_FORMAT
4686 G1PPRL_BYTE_FORMAT
4687 G1PPRL_DOUBLE_FORMAT
4688 G1PPRL_BYTE_FORMAT
4689 G1PPRL_BYTE_FORMAT,
4690 type, p2i(bottom), p2i(end),
4691 used_bytes, prev_live_bytes, next_live_bytes, gc_eff,
4692 remset_bytes, strong_code_roots_bytes);
4693
4694 return false;
4695 }
4696
4697 G1PrintRegionLivenessInfoClosure::~G1PrintRegionLivenessInfoClosure() {
4698 // add static memory usages to remembered set sizes
4699 _total_remset_bytes += HeapRegionRemSet::fl_mem_size() + HeapRegionRemSet::static_mem_size();
4700 // Print the footer of the output.
4701 _out->print_cr(G1PPRL_LINE_PREFIX);
4702 _out->print_cr(G1PPRL_LINE_PREFIX
4703 " SUMMARY"
4704 G1PPRL_SUM_MB_FORMAT("capacity")
4705 G1PPRL_SUM_MB_PERC_FORMAT("used")
4706 G1PPRL_SUM_MB_PERC_FORMAT("prev-live")
4707 G1PPRL_SUM_MB_PERC_FORMAT("next-live")
4708 G1PPRL_SUM_MB_FORMAT("remset")
4709 G1PPRL_SUM_MB_FORMAT("code-roots"),
4710 bytes_to_mb(_total_capacity_bytes),
4711 bytes_to_mb(_total_used_bytes),
4712 perc(_total_used_bytes, _total_capacity_bytes),
4713 bytes_to_mb(_total_prev_live_bytes),
4714 perc(_total_prev_live_bytes, _total_capacity_bytes),
4715 bytes_to_mb(_total_next_live_bytes),
4716 perc(_total_next_live_bytes, _total_capacity_bytes),
4717 bytes_to_mb(_total_remset_bytes),
4718 bytes_to_mb(_total_strong_code_roots_bytes));
4719 _out->cr();
4720 }