1 /*
2 * Copyright (c) 2001, 2015, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation.
8 *
9 * This code is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 * version 2 for more details (a copy is included in the LICENSE file that
13 * accompanied this code).
14 *
15 * You should have received a copy of the GNU General Public License version
16 * 2 along with this work; if not, write to the Free Software Foundation,
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18 *
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20 * or visit www.oracle.com if you need additional information or have any
21 * questions.
22 *
23 */
24
25 #include "precompiled.hpp"
26 #include "gc/g1/concurrentG1Refine.hpp"
27 #include "gc/g1/concurrentMark.hpp"
28 #include "gc/g1/concurrentMarkThread.inline.hpp"
29 #include "gc/g1/g1CollectedHeap.inline.hpp"
30 #include "gc/g1/g1CollectorPolicy.hpp"
31 #include "gc/g1/g1IHOPControl.hpp"
32 #include "gc/g1/g1GCPhaseTimes.hpp"
33 #include "gc/g1/heapRegion.inline.hpp"
34 #include "gc/g1/heapRegionRemSet.hpp"
35 #include "gc/shared/gcPolicyCounters.hpp"
36 #include "runtime/arguments.hpp"
37 #include "runtime/java.hpp"
38 #include "runtime/mutexLocker.hpp"
39 #include "utilities/debug.hpp"
40 #include "utilities/pair.hpp"
41
42 // Different defaults for different number of GC threads
43 // They were chosen by running GCOld and SPECjbb on debris with different
44 // numbers of GC threads and choosing them based on the results
45
46 // all the same
47 static double rs_length_diff_defaults[] = {
48 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0
49 };
50
51 static double cost_per_card_ms_defaults[] = {
52 0.01, 0.005, 0.005, 0.003, 0.003, 0.002, 0.002, 0.0015
53 };
54
55 // all the same
56 static double young_cards_per_entry_ratio_defaults[] = {
57 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0
58 };
59
60 static double cost_per_entry_ms_defaults[] = {
61 0.015, 0.01, 0.01, 0.008, 0.008, 0.0055, 0.0055, 0.005
62 };
63
64 static double cost_per_byte_ms_defaults[] = {
65 0.00006, 0.00003, 0.00003, 0.000015, 0.000015, 0.00001, 0.00001, 0.000009
66 };
67
68 // these should be pretty consistent
69 static double constant_other_time_ms_defaults[] = {
70 5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0, 5.0
71 };
72
73
74 static double young_other_cost_per_region_ms_defaults[] = {
75 0.3, 0.2, 0.2, 0.15, 0.15, 0.12, 0.12, 0.1
76 };
77
78 static double non_young_other_cost_per_region_ms_defaults[] = {
79 1.0, 0.7, 0.7, 0.5, 0.5, 0.42, 0.42, 0.30
80 };
81
82 G1CollectorPolicy::G1CollectorPolicy() :
83 _predictor(G1ConfidencePercent / 100.0),
84 _parallel_gc_threads(ParallelGCThreads),
85
86 _recent_gc_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
87 _stop_world_start(0.0),
88
89 _concurrent_mark_remark_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
90 _concurrent_mark_cleanup_times_ms(new TruncatedSeq(NumPrevPausesForHeuristics)),
91
92 _alloc_rate_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
93 _prev_collection_pause_end_ms(0.0),
94 _rs_length_diff_seq(new TruncatedSeq(TruncatedSeqLength)),
95 _cost_per_card_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
96 _cost_scan_hcc_seq(new TruncatedSeq(TruncatedSeqLength)),
97 _young_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
98 _mixed_cards_per_entry_ratio_seq(new TruncatedSeq(TruncatedSeqLength)),
99 _cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
100 _mixed_cost_per_entry_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
101 _cost_per_byte_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
102 _cost_per_byte_ms_during_cm_seq(new TruncatedSeq(TruncatedSeqLength)),
103 _constant_other_time_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
104 _young_other_cost_per_region_ms_seq(new TruncatedSeq(TruncatedSeqLength)),
105 _non_young_other_cost_per_region_ms_seq(
106 new TruncatedSeq(TruncatedSeqLength)),
107
108 _pending_cards_seq(new TruncatedSeq(TruncatedSeqLength)),
109 _rs_lengths_seq(new TruncatedSeq(TruncatedSeqLength)),
110
111 _pause_time_target_ms((double) MaxGCPauseMillis),
112
113 _recent_prev_end_times_for_all_gcs_sec(
114 new TruncatedSeq(NumPrevPausesForHeuristics)),
115
116 _recent_avg_pause_time_ratio(0.0),
117 _rs_lengths_prediction(0),
118 _max_survivor_regions(0),
119
120 _eden_used_bytes_before_gc(0),
121 _survivor_used_bytes_before_gc(0),
122 _old_used_bytes_before_gc(0),
123 _humongous_used_bytes_before_gc(0),
124 _heap_used_bytes_before_gc(0),
125 _metaspace_used_bytes_before_gc(0),
126 _eden_capacity_bytes_before_gc(0),
127 _heap_capacity_bytes_before_gc(0),
128
129 _eden_cset_region_length(0),
130 _survivor_cset_region_length(0),
131 _old_cset_region_length(0),
132
133 _collection_set(NULL),
134 _collection_set_bytes_used_before(0),
135
136 // Incremental CSet attributes
137 _inc_cset_build_state(Inactive),
138 _inc_cset_head(NULL),
139 _inc_cset_tail(NULL),
140 _inc_cset_bytes_used_before(0),
141 _inc_cset_max_finger(NULL),
142 _inc_cset_recorded_rs_lengths(0),
143 _inc_cset_recorded_rs_lengths_diffs(0),
144 _inc_cset_predicted_elapsed_time_ms(0.0),
145 _inc_cset_predicted_elapsed_time_ms_diffs(0.0),
146
147 // add here any more surv rate groups
148 _recorded_survivor_regions(0),
149 _recorded_survivor_head(NULL),
150 _recorded_survivor_tail(NULL),
151 _survivors_age_table(true),
152
153 _gc_overhead_perc(0.0),
154
155 _bytes_allocated_in_old_since_last_gc(0),
156 _ihop_control(NULL),
157 _initial_mark_to_mixed() {
158
159 // SurvRateGroups below must be initialized after the predictor because they
160 // indirectly use it through this object passed to their constructor.
161 _short_lived_surv_rate_group =
162 new SurvRateGroup(&_predictor, "Short Lived", G1YoungSurvRateNumRegionsSummary);
163 _survivor_surv_rate_group =
164 new SurvRateGroup(&_predictor, "Survivor", G1YoungSurvRateNumRegionsSummary);
165
166 // Set up the region size and associated fields. Given that the
167 // policy is created before the heap, we have to set this up here,
168 // so it's done as soon as possible.
169
170 // It would have been natural to pass initial_heap_byte_size() and
171 // max_heap_byte_size() to setup_heap_region_size() but those have
172 // not been set up at this point since they should be aligned with
173 // the region size. So, there is a circular dependency here. We base
174 // the region size on the heap size, but the heap size should be
175 // aligned with the region size. To get around this we use the
176 // unaligned values for the heap.
177 HeapRegion::setup_heap_region_size(InitialHeapSize, MaxHeapSize);
178 HeapRegionRemSet::setup_remset_size();
179
180 _recent_prev_end_times_for_all_gcs_sec->add(os::elapsedTime());
181 _prev_collection_pause_end_ms = os::elapsedTime() * 1000.0;
182 clear_ratio_check_data();
183
184 _phase_times = new G1GCPhaseTimes(_parallel_gc_threads);
185
186 int index = MIN2(_parallel_gc_threads - 1, 7);
187
188 _rs_length_diff_seq->add(rs_length_diff_defaults[index]);
189 _cost_per_card_ms_seq->add(cost_per_card_ms_defaults[index]);
190 _cost_scan_hcc_seq->add(0.0);
191 _young_cards_per_entry_ratio_seq->add(
192 young_cards_per_entry_ratio_defaults[index]);
193 _cost_per_entry_ms_seq->add(cost_per_entry_ms_defaults[index]);
194 _cost_per_byte_ms_seq->add(cost_per_byte_ms_defaults[index]);
195 _constant_other_time_ms_seq->add(constant_other_time_ms_defaults[index]);
196 _young_other_cost_per_region_ms_seq->add(
197 young_other_cost_per_region_ms_defaults[index]);
198 _non_young_other_cost_per_region_ms_seq->add(
199 non_young_other_cost_per_region_ms_defaults[index]);
200
201 // Below, we might need to calculate the pause time target based on
202 // the pause interval. When we do so we are going to give G1 maximum
203 // flexibility and allow it to do pauses when it needs to. So, we'll
204 // arrange that the pause interval to be pause time target + 1 to
205 // ensure that a) the pause time target is maximized with respect to
206 // the pause interval and b) we maintain the invariant that pause
207 // time target < pause interval. If the user does not want this
208 // maximum flexibility, they will have to set the pause interval
209 // explicitly.
210
211 // First make sure that, if either parameter is set, its value is
212 // reasonable.
213 if (!FLAG_IS_DEFAULT(MaxGCPauseMillis)) {
214 if (MaxGCPauseMillis < 1) {
215 vm_exit_during_initialization("MaxGCPauseMillis should be "
216 "greater than 0");
217 }
218 }
219 if (!FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
220 if (GCPauseIntervalMillis < 1) {
221 vm_exit_during_initialization("GCPauseIntervalMillis should be "
222 "greater than 0");
223 }
224 }
225
226 // Then, if the pause time target parameter was not set, set it to
227 // the default value.
228 if (FLAG_IS_DEFAULT(MaxGCPauseMillis)) {
229 if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
230 // The default pause time target in G1 is 200ms
231 FLAG_SET_DEFAULT(MaxGCPauseMillis, 200);
232 } else {
233 // We do not allow the pause interval to be set without the
234 // pause time target
235 vm_exit_during_initialization("GCPauseIntervalMillis cannot be set "
236 "without setting MaxGCPauseMillis");
237 }
238 }
239
240 // Then, if the interval parameter was not set, set it according to
241 // the pause time target (this will also deal with the case when the
242 // pause time target is the default value).
243 if (FLAG_IS_DEFAULT(GCPauseIntervalMillis)) {
244 FLAG_SET_DEFAULT(GCPauseIntervalMillis, MaxGCPauseMillis + 1);
245 }
246
247 // Finally, make sure that the two parameters are consistent.
248 if (MaxGCPauseMillis >= GCPauseIntervalMillis) {
249 char buffer[256];
250 jio_snprintf(buffer, 256,
251 "MaxGCPauseMillis (%u) should be less than "
252 "GCPauseIntervalMillis (%u)",
253 MaxGCPauseMillis, GCPauseIntervalMillis);
254 vm_exit_during_initialization(buffer);
255 }
256
257 double max_gc_time = (double) MaxGCPauseMillis / 1000.0;
258 double time_slice = (double) GCPauseIntervalMillis / 1000.0;
259 _mmu_tracker = new G1MMUTrackerQueue(time_slice, max_gc_time);
260
261 // start conservatively (around 50ms is about right)
262 _concurrent_mark_remark_times_ms->add(0.05);
263 _concurrent_mark_cleanup_times_ms->add(0.20);
264 _tenuring_threshold = MaxTenuringThreshold;
265
266 assert(GCTimeRatio > 0,
267 "we should have set it to a default value set_g1_gc_flags() "
268 "if a user set it to 0");
269 _gc_overhead_perc = 100.0 * (1.0 / (1.0 + GCTimeRatio));
270
271 uintx reserve_perc = G1ReservePercent;
272 // Put an artificial ceiling on this so that it's not set to a silly value.
273 if (reserve_perc > 50) {
274 reserve_perc = 50;
275 warning("G1ReservePercent is set to a value that is too large, "
276 "it's been updated to " UINTX_FORMAT, reserve_perc);
277 }
278 _reserve_factor = (double) reserve_perc / 100.0;
279 // This will be set when the heap is expanded
280 // for the first time during initialization.
281 _reserve_regions = 0;
282
283 _cset_chooser = new CollectionSetChooser();
284 }
285
286 G1CollectorPolicy::~G1CollectorPolicy() {
287 delete _ihop_control;
288 }
289
290 double G1CollectorPolicy::get_new_prediction(TruncatedSeq const* seq) const {
291 return _predictor.get_new_prediction(seq);
292 }
293
294 void G1CollectorPolicy::initialize_alignments() {
295 _space_alignment = HeapRegion::GrainBytes;
296 size_t card_table_alignment = CardTableRS::ct_max_alignment_constraint();
297 size_t page_size = UseLargePages ? os::large_page_size() : os::vm_page_size();
298 _heap_alignment = MAX3(card_table_alignment, _space_alignment, page_size);
299 }
300
301 void G1CollectorPolicy::initialize_flags() {
302 if (G1HeapRegionSize != HeapRegion::GrainBytes) {
303 FLAG_SET_ERGO(size_t, G1HeapRegionSize, HeapRegion::GrainBytes);
304 }
305
306 if (SurvivorRatio < 1) {
307 vm_exit_during_initialization("Invalid survivor ratio specified");
308 }
309 CollectorPolicy::initialize_flags();
310 _young_gen_sizer = new G1YoungGenSizer(); // Must be after call to initialize_flags
311 }
312
313 void G1CollectorPolicy::post_heap_initialize() {
314 uintx max_regions = G1CollectedHeap::heap()->max_regions();
315 size_t max_young_size = (size_t)_young_gen_sizer->max_young_length(max_regions) * HeapRegion::GrainBytes;
316 if (max_young_size != MaxNewSize) {
317 FLAG_SET_ERGO(size_t, MaxNewSize, max_young_size);
318 }
319
320 _ihop_control = create_ihop_control();
321 }
322
323 G1CollectorState* G1CollectorPolicy::collector_state() const { return _g1->collector_state(); }
324
325 G1YoungGenSizer::G1YoungGenSizer() : _sizer_kind(SizerDefaults), _adaptive_size(true),
326 _min_desired_young_length(0), _max_desired_young_length(0) {
327 if (FLAG_IS_CMDLINE(NewRatio)) {
328 if (FLAG_IS_CMDLINE(NewSize) || FLAG_IS_CMDLINE(MaxNewSize)) {
329 warning("-XX:NewSize and -XX:MaxNewSize override -XX:NewRatio");
330 } else {
331 _sizer_kind = SizerNewRatio;
332 _adaptive_size = false;
333 return;
334 }
335 }
336
337 if (NewSize > MaxNewSize) {
338 if (FLAG_IS_CMDLINE(MaxNewSize)) {
339 warning("NewSize (" SIZE_FORMAT "k) is greater than the MaxNewSize (" SIZE_FORMAT "k). "
340 "A new max generation size of " SIZE_FORMAT "k will be used.",
341 NewSize/K, MaxNewSize/K, NewSize/K);
342 }
343 MaxNewSize = NewSize;
344 }
345
346 if (FLAG_IS_CMDLINE(NewSize)) {
347 _min_desired_young_length = MAX2((uint) (NewSize / HeapRegion::GrainBytes),
348 1U);
349 if (FLAG_IS_CMDLINE(MaxNewSize)) {
350 _max_desired_young_length =
351 MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes),
352 1U);
353 _sizer_kind = SizerMaxAndNewSize;
354 _adaptive_size = _min_desired_young_length == _max_desired_young_length;
355 } else {
356 _sizer_kind = SizerNewSizeOnly;
357 }
358 } else if (FLAG_IS_CMDLINE(MaxNewSize)) {
359 _max_desired_young_length =
360 MAX2((uint) (MaxNewSize / HeapRegion::GrainBytes),
361 1U);
362 _sizer_kind = SizerMaxNewSizeOnly;
363 }
364 }
365
366 uint G1YoungGenSizer::calculate_default_min_length(uint new_number_of_heap_regions) {
367 uint default_value = (new_number_of_heap_regions * G1NewSizePercent) / 100;
368 return MAX2(1U, default_value);
369 }
370
371 uint G1YoungGenSizer::calculate_default_max_length(uint new_number_of_heap_regions) {
372 uint default_value = (new_number_of_heap_regions * G1MaxNewSizePercent) / 100;
373 return MAX2(1U, default_value);
374 }
375
376 void G1YoungGenSizer::recalculate_min_max_young_length(uint number_of_heap_regions, uint* min_young_length, uint* max_young_length) {
377 assert(number_of_heap_regions > 0, "Heap must be initialized");
378
379 switch (_sizer_kind) {
380 case SizerDefaults:
381 *min_young_length = calculate_default_min_length(number_of_heap_regions);
382 *max_young_length = calculate_default_max_length(number_of_heap_regions);
383 break;
384 case SizerNewSizeOnly:
385 *max_young_length = calculate_default_max_length(number_of_heap_regions);
386 *max_young_length = MAX2(*min_young_length, *max_young_length);
387 break;
388 case SizerMaxNewSizeOnly:
389 *min_young_length = calculate_default_min_length(number_of_heap_regions);
390 *min_young_length = MIN2(*min_young_length, *max_young_length);
391 break;
392 case SizerMaxAndNewSize:
393 // Do nothing. Values set on the command line, don't update them at runtime.
394 break;
395 case SizerNewRatio:
396 *min_young_length = number_of_heap_regions / (NewRatio + 1);
397 *max_young_length = *min_young_length;
398 break;
399 default:
400 ShouldNotReachHere();
401 }
402
403 assert(*min_young_length <= *max_young_length, "Invalid min/max young gen size values");
404 }
405
406 uint G1YoungGenSizer::max_young_length(uint number_of_heap_regions) {
407 // We need to pass the desired values because recalculation may not update these
408 // values in some cases.
409 uint temp = _min_desired_young_length;
410 uint result = _max_desired_young_length;
411 recalculate_min_max_young_length(number_of_heap_regions, &temp, &result);
412 return result;
413 }
414
415 void G1YoungGenSizer::heap_size_changed(uint new_number_of_heap_regions) {
416 recalculate_min_max_young_length(new_number_of_heap_regions, &_min_desired_young_length,
417 &_max_desired_young_length);
418 }
419
420 void G1CollectorPolicy::init() {
421 // Set aside an initial future to_space.
422 _g1 = G1CollectedHeap::heap();
423
424 assert(Heap_lock->owned_by_self(), "Locking discipline.");
425
426 initialize_gc_policy_counters();
427
428 if (adaptive_young_list_length()) {
429 _young_list_fixed_length = 0;
430 } else {
431 _young_list_fixed_length = _young_gen_sizer->min_desired_young_length();
432 }
433 _free_regions_at_end_of_collection = _g1->num_free_regions();
434
435 update_young_list_max_and_target_length();
436 // We may immediately start allocating regions and placing them on the
437 // collection set list. Initialize the per-collection set info
438 start_incremental_cset_building();
439 }
440
441 void G1CollectorPolicy::note_gc_start(uint num_active_workers) {
442 phase_times()->note_gc_start(num_active_workers);
443 }
444
445 // Create the jstat counters for the policy.
446 void G1CollectorPolicy::initialize_gc_policy_counters() {
447 _gc_policy_counters = new GCPolicyCounters("GarbageFirst", 1, 3);
448 }
449
450 bool G1CollectorPolicy::predict_will_fit(uint young_length,
451 double base_time_ms,
452 uint base_free_regions,
453 double target_pause_time_ms) const {
454 if (young_length >= base_free_regions) {
455 // end condition 1: not enough space for the young regions
456 return false;
457 }
458
459 double accum_surv_rate = accum_yg_surv_rate_pred((int) young_length - 1);
460 size_t bytes_to_copy =
461 (size_t) (accum_surv_rate * (double) HeapRegion::GrainBytes);
462 double copy_time_ms = predict_object_copy_time_ms(bytes_to_copy);
463 double young_other_time_ms = predict_young_other_time_ms(young_length);
464 double pause_time_ms = base_time_ms + copy_time_ms + young_other_time_ms;
465 if (pause_time_ms > target_pause_time_ms) {
466 // end condition 2: prediction is over the target pause time
467 return false;
468 }
469
470 size_t free_bytes = (base_free_regions - young_length) * HeapRegion::GrainBytes;
471
472 // When copying, we will likely need more bytes free than is live in the region.
473 // Add some safety margin to factor in the confidence of our guess, and the
474 // natural expected waste.
475 // (100.0 / G1ConfidencePercent) is a scale factor that expresses the uncertainty
476 // of the calculation: the lower the confidence, the more headroom.
477 // (100 + TargetPLABWastePct) represents the increase in expected bytes during
478 // copying due to anticipated waste in the PLABs.
479 double safety_factor = (100.0 / G1ConfidencePercent) * (100 + TargetPLABWastePct) / 100.0;
480 size_t expected_bytes_to_copy = safety_factor * bytes_to_copy;
481
482 if (expected_bytes_to_copy > free_bytes) {
483 // end condition 3: out-of-space
484 return false;
485 }
486
487 // success!
488 return true;
489 }
490
491 void G1CollectorPolicy::record_new_heap_size(uint new_number_of_regions) {
492 // re-calculate the necessary reserve
493 double reserve_regions_d = (double) new_number_of_regions * _reserve_factor;
494 // We use ceiling so that if reserve_regions_d is > 0.0 (but
495 // smaller than 1.0) we'll get 1.
496 _reserve_regions = (uint) ceil(reserve_regions_d);
497
498 _young_gen_sizer->heap_size_changed(new_number_of_regions);
499 }
500
501 uint G1CollectorPolicy::calculate_young_list_desired_min_length(
502 uint base_min_length) const {
503 uint desired_min_length = 0;
504 if (adaptive_young_list_length()) {
505 if (_alloc_rate_ms_seq->num() > 3) {
506 double now_sec = os::elapsedTime();
507 double when_ms = _mmu_tracker->when_max_gc_sec(now_sec) * 1000.0;
508 double alloc_rate_ms = predict_alloc_rate_ms();
509 desired_min_length = (uint) ceil(alloc_rate_ms * when_ms);
510 } else {
511 // otherwise we don't have enough info to make the prediction
512 }
513 }
514 desired_min_length += base_min_length;
515 // make sure we don't go below any user-defined minimum bound
516 return MAX2(_young_gen_sizer->min_desired_young_length(), desired_min_length);
517 }
518
519 uint G1CollectorPolicy::calculate_young_list_desired_max_length() const {
520 // Here, we might want to also take into account any additional
521 // constraints (i.e., user-defined minimum bound). Currently, we
522 // effectively don't set this bound.
523 return _young_gen_sizer->max_desired_young_length();
524 }
525
526 uint G1CollectorPolicy::update_young_list_max_and_target_length() {
527 return update_young_list_max_and_target_length(get_new_prediction(_rs_lengths_seq));
528 }
529
530 uint G1CollectorPolicy::update_young_list_max_and_target_length(size_t rs_lengths) {
531 uint unbounded_target_length = update_young_list_target_length(rs_lengths);
532 update_max_gc_locker_expansion();
533 return unbounded_target_length;
534 }
535
536 uint G1CollectorPolicy::update_young_list_target_length(size_t rs_lengths) {
537 YoungTargetLengths young_lengths = young_list_target_lengths(rs_lengths);
538 _young_list_target_length = young_lengths.first;
539 return young_lengths.second;
540 }
541
542 G1CollectorPolicy::YoungTargetLengths G1CollectorPolicy::young_list_target_lengths(size_t rs_lengths) const {
543 YoungTargetLengths result;
544
545 // Calculate the absolute and desired min bounds first.
546
547 // This is how many young regions we already have (currently: the survivors).
548 uint base_min_length = recorded_survivor_regions();
549 uint desired_min_length = calculate_young_list_desired_min_length(base_min_length);
550 // This is the absolute minimum young length. Ensure that we
551 // will at least have one eden region available for allocation.
552 uint absolute_min_length = base_min_length + MAX2(_g1->young_list()->eden_length(), (uint)1);
553 // If we shrank the young list target it should not shrink below the current size.
554 desired_min_length = MAX2(desired_min_length, absolute_min_length);
555 // Calculate the absolute and desired max bounds.
556
557 uint desired_max_length = calculate_young_list_desired_max_length();
558
559 uint young_list_target_length = 0;
560 if (adaptive_young_list_length()) {
561 if (collector_state()->gcs_are_young()) {
562 young_list_target_length =
563 calculate_young_list_target_length(rs_lengths,
564 base_min_length,
565 desired_min_length,
566 desired_max_length);
567 } else {
568 // Don't calculate anything and let the code below bound it to
569 // the desired_min_length, i.e., do the next GC as soon as
570 // possible to maximize how many old regions we can add to it.
571 }
572 } else {
573 // The user asked for a fixed young gen so we'll fix the young gen
574 // whether the next GC is young or mixed.
575 young_list_target_length = _young_list_fixed_length;
576 }
577
578 result.second = young_list_target_length;
579
580 // We will try our best not to "eat" into the reserve.
581 uint absolute_max_length = 0;
582 if (_free_regions_at_end_of_collection > _reserve_regions) {
583 absolute_max_length = _free_regions_at_end_of_collection - _reserve_regions;
584 }
585 if (desired_max_length > absolute_max_length) {
586 desired_max_length = absolute_max_length;
587 }
588
589 // Make sure we don't go over the desired max length, nor under the
590 // desired min length. In case they clash, desired_min_length wins
591 // which is why that test is second.
592 if (young_list_target_length > desired_max_length) {
593 young_list_target_length = desired_max_length;
594 }
595 if (young_list_target_length < desired_min_length) {
596 young_list_target_length = desired_min_length;
597 }
598
599 assert(young_list_target_length > recorded_survivor_regions(),
600 "we should be able to allocate at least one eden region");
601 assert(young_list_target_length >= absolute_min_length, "post-condition");
602
603 result.first = young_list_target_length;
604 return result;
605 }
606
607 uint
608 G1CollectorPolicy::calculate_young_list_target_length(size_t rs_lengths,
609 uint base_min_length,
610 uint desired_min_length,
611 uint desired_max_length) const {
612 assert(adaptive_young_list_length(), "pre-condition");
613 assert(collector_state()->gcs_are_young(), "only call this for young GCs");
614
615 // In case some edge-condition makes the desired max length too small...
616 if (desired_max_length <= desired_min_length) {
617 return desired_min_length;
618 }
619
620 // We'll adjust min_young_length and max_young_length not to include
621 // the already allocated young regions (i.e., so they reflect the
622 // min and max eden regions we'll allocate). The base_min_length
623 // will be reflected in the predictions by the
624 // survivor_regions_evac_time prediction.
625 assert(desired_min_length > base_min_length, "invariant");
626 uint min_young_length = desired_min_length - base_min_length;
627 assert(desired_max_length > base_min_length, "invariant");
628 uint max_young_length = desired_max_length - base_min_length;
629
630 double target_pause_time_ms = _mmu_tracker->max_gc_time() * 1000.0;
631 double survivor_regions_evac_time = predict_survivor_regions_evac_time();
632 size_t pending_cards = (size_t) get_new_prediction(_pending_cards_seq);
633 size_t adj_rs_lengths = rs_lengths + predict_rs_length_diff();
634 size_t scanned_cards = predict_young_card_num(adj_rs_lengths);
635 double base_time_ms =
636 predict_base_elapsed_time_ms(pending_cards, scanned_cards) +
637 survivor_regions_evac_time;
638 uint available_free_regions = _free_regions_at_end_of_collection;
639 uint base_free_regions = 0;
640 if (available_free_regions > _reserve_regions) {
641 base_free_regions = available_free_regions - _reserve_regions;
642 }
643
644 // Here, we will make sure that the shortest young length that
645 // makes sense fits within the target pause time.
646
647 if (predict_will_fit(min_young_length, base_time_ms,
648 base_free_regions, target_pause_time_ms)) {
649 // The shortest young length will fit into the target pause time;
650 // we'll now check whether the absolute maximum number of young
651 // regions will fit in the target pause time. If not, we'll do
652 // a binary search between min_young_length and max_young_length.
653 if (predict_will_fit(max_young_length, base_time_ms,
654 base_free_regions, target_pause_time_ms)) {
655 // The maximum young length will fit into the target pause time.
656 // We are done so set min young length to the maximum length (as
657 // the result is assumed to be returned in min_young_length).
658 min_young_length = max_young_length;
659 } else {
660 // The maximum possible number of young regions will not fit within
661 // the target pause time so we'll search for the optimal
662 // length. The loop invariants are:
663 //
664 // min_young_length < max_young_length
665 // min_young_length is known to fit into the target pause time
666 // max_young_length is known not to fit into the target pause time
667 //
668 // Going into the loop we know the above hold as we've just
669 // checked them. Every time around the loop we check whether
670 // the middle value between min_young_length and
671 // max_young_length fits into the target pause time. If it
672 // does, it becomes the new min. If it doesn't, it becomes
673 // the new max. This way we maintain the loop invariants.
674
675 assert(min_young_length < max_young_length, "invariant");
676 uint diff = (max_young_length - min_young_length) / 2;
677 while (diff > 0) {
678 uint young_length = min_young_length + diff;
679 if (predict_will_fit(young_length, base_time_ms,
680 base_free_regions, target_pause_time_ms)) {
681 min_young_length = young_length;
682 } else {
683 max_young_length = young_length;
684 }
685 assert(min_young_length < max_young_length, "invariant");
686 diff = (max_young_length - min_young_length) / 2;
687 }
688 // The results is min_young_length which, according to the
689 // loop invariants, should fit within the target pause time.
690
691 // These are the post-conditions of the binary search above:
692 assert(min_young_length < max_young_length,
693 "otherwise we should have discovered that max_young_length "
694 "fits into the pause target and not done the binary search");
695 assert(predict_will_fit(min_young_length, base_time_ms,
696 base_free_regions, target_pause_time_ms),
697 "min_young_length, the result of the binary search, should "
698 "fit into the pause target");
699 assert(!predict_will_fit(min_young_length + 1, base_time_ms,
700 base_free_regions, target_pause_time_ms),
701 "min_young_length, the result of the binary search, should be "
702 "optimal, so no larger length should fit into the pause target");
703 }
704 } else {
705 // Even the minimum length doesn't fit into the pause time
706 // target, return it as the result nevertheless.
707 }
708 return base_min_length + min_young_length;
709 }
710
711 double G1CollectorPolicy::predict_survivor_regions_evac_time() const {
712 double survivor_regions_evac_time = 0.0;
713 for (HeapRegion * r = _recorded_survivor_head;
714 r != NULL && r != _recorded_survivor_tail->get_next_young_region();
715 r = r->get_next_young_region()) {
716 survivor_regions_evac_time += predict_region_elapsed_time_ms(r, collector_state()->gcs_are_young());
717 }
718 return survivor_regions_evac_time;
719 }
720
721 void G1CollectorPolicy::revise_young_list_target_length_if_necessary() {
722 guarantee( adaptive_young_list_length(), "should not call this otherwise" );
723
724 size_t rs_lengths = _g1->young_list()->sampled_rs_lengths();
725 if (rs_lengths > _rs_lengths_prediction) {
726 // add 10% to avoid having to recalculate often
727 size_t rs_lengths_prediction = rs_lengths * 1100 / 1000;
728 update_rs_lengths_prediction(rs_lengths_prediction);
729
730 update_young_list_max_and_target_length(rs_lengths_prediction);
731 }
732 }
733
734 void G1CollectorPolicy::update_rs_lengths_prediction() {
735 update_rs_lengths_prediction(get_new_prediction(_rs_lengths_seq));
736 }
737
738 void G1CollectorPolicy::update_rs_lengths_prediction(size_t prediction) {
739 if (collector_state()->gcs_are_young() && adaptive_young_list_length()) {
740 _rs_lengths_prediction = prediction;
741 }
742 }
743
744 HeapWord* G1CollectorPolicy::mem_allocate_work(size_t size,
745 bool is_tlab,
746 bool* gc_overhead_limit_was_exceeded) {
747 guarantee(false, "Not using this policy feature yet.");
748 return NULL;
749 }
750
751 // This method controls how a collector handles one or more
752 // of its generations being fully allocated.
753 HeapWord* G1CollectorPolicy::satisfy_failed_allocation(size_t size,
754 bool is_tlab) {
755 guarantee(false, "Not using this policy feature yet.");
756 return NULL;
757 }
758
759
760 #ifndef PRODUCT
761 bool G1CollectorPolicy::verify_young_ages() {
762 HeapRegion* head = _g1->young_list()->first_region();
763 return
764 verify_young_ages(head, _short_lived_surv_rate_group);
765 // also call verify_young_ages on any additional surv rate groups
766 }
767
768 bool
769 G1CollectorPolicy::verify_young_ages(HeapRegion* head,
770 SurvRateGroup *surv_rate_group) {
771 guarantee( surv_rate_group != NULL, "pre-condition" );
772
773 const char* name = surv_rate_group->name();
774 bool ret = true;
775 int prev_age = -1;
776
777 for (HeapRegion* curr = head;
778 curr != NULL;
779 curr = curr->get_next_young_region()) {
780 SurvRateGroup* group = curr->surv_rate_group();
781 if (group == NULL && !curr->is_survivor()) {
782 log_info(gc, verify)("## %s: encountered NULL surv_rate_group", name);
783 ret = false;
784 }
785
786 if (surv_rate_group == group) {
787 int age = curr->age_in_surv_rate_group();
788
789 if (age < 0) {
790 log_info(gc, verify)("## %s: encountered negative age", name);
791 ret = false;
792 }
793
794 if (age <= prev_age) {
795 log_info(gc, verify)("## %s: region ages are not strictly increasing (%d, %d)", name, age, prev_age);
796 ret = false;
797 }
798 prev_age = age;
799 }
800 }
801
802 return ret;
803 }
804 #endif // PRODUCT
805
806 void G1CollectorPolicy::record_full_collection_start() {
807 _full_collection_start_sec = os::elapsedTime();
808 record_heap_size_info_at_start(true /* full */);
809 // Release the future to-space so that it is available for compaction into.
810 collector_state()->set_full_collection(true);
811 }
812
813 void G1CollectorPolicy::record_full_collection_end() {
814 // Consider this like a collection pause for the purposes of allocation
815 // since last pause.
816 double end_sec = os::elapsedTime();
817 double full_gc_time_sec = end_sec - _full_collection_start_sec;
818 double full_gc_time_ms = full_gc_time_sec * 1000.0;
819
820 _trace_old_gen_time_data.record_full_collection(full_gc_time_ms);
821
822 update_recent_gc_times(end_sec, full_gc_time_ms);
823
824 collector_state()->set_full_collection(false);
825
826 // "Nuke" the heuristics that control the young/mixed GC
827 // transitions and make sure we start with young GCs after the Full GC.
828 collector_state()->set_gcs_are_young(true);
829 collector_state()->set_last_young_gc(false);
830 collector_state()->set_initiate_conc_mark_if_possible(need_to_start_conc_mark("end of Full GC", 0));
831 collector_state()->set_during_initial_mark_pause(false);
832 collector_state()->set_in_marking_window(false);
833 collector_state()->set_in_marking_window_im(false);
834
835 _short_lived_surv_rate_group->start_adding_regions();
836 // also call this on any additional surv rate groups
837
838 record_survivor_regions(0, NULL, NULL);
839
840 _free_regions_at_end_of_collection = _g1->num_free_regions();
841 // Reset survivors SurvRateGroup.
842 _survivor_surv_rate_group->reset();
843 update_young_list_max_and_target_length();
844 update_rs_lengths_prediction();
845 cset_chooser()->clear();
846
847 _bytes_allocated_in_old_since_last_gc = 0;
848
849 record_pause(FullGC, _full_collection_start_sec, end_sec);
850 }
851
852 void G1CollectorPolicy::record_stop_world_start() {
853 _stop_world_start = os::elapsedTime();
854 }
855
856 void G1CollectorPolicy::record_collection_pause_start(double start_time_sec) {
857 // We only need to do this here as the policy will only be applied
858 // to the GC we're about to start. so, no point is calculating this
859 // every time we calculate / recalculate the target young length.
860 update_survivors_policy();
861
862 assert(_g1->used() == _g1->recalculate_used(),
863 "sanity, used: " SIZE_FORMAT " recalculate_used: " SIZE_FORMAT,
864 _g1->used(), _g1->recalculate_used());
865
866 double s_w_t_ms = (start_time_sec - _stop_world_start) * 1000.0;
867 _trace_young_gen_time_data.record_start_collection(s_w_t_ms);
868 _stop_world_start = 0.0;
869
870 record_heap_size_info_at_start(false /* full */);
871
872 phase_times()->record_cur_collection_start_sec(start_time_sec);
873 _pending_cards = _g1->pending_card_num();
874
875 _collection_set_bytes_used_before = 0;
876 _bytes_copied_during_gc = 0;
877
878 collector_state()->set_last_gc_was_young(false);
879
880 // do that for any other surv rate groups
881 _short_lived_surv_rate_group->stop_adding_regions();
882 _survivors_age_table.clear();
883
884 assert( verify_young_ages(), "region age verification" );
885 }
886
887 void G1CollectorPolicy::record_concurrent_mark_init_end(double
888 mark_init_elapsed_time_ms) {
889 collector_state()->set_during_marking(true);
890 assert(!collector_state()->initiate_conc_mark_if_possible(), "we should have cleared it by now");
891 collector_state()->set_during_initial_mark_pause(false);
892 }
893
894 void G1CollectorPolicy::record_concurrent_mark_remark_start() {
895 _mark_remark_start_sec = os::elapsedTime();
896 collector_state()->set_during_marking(false);
897 }
898
899 void G1CollectorPolicy::record_concurrent_mark_remark_end() {
900 double end_time_sec = os::elapsedTime();
901 double elapsed_time_ms = (end_time_sec - _mark_remark_start_sec)*1000.0;
902 _concurrent_mark_remark_times_ms->add(elapsed_time_ms);
903 _prev_collection_pause_end_ms += elapsed_time_ms;
904
905 record_pause(Remark, _mark_remark_start_sec, end_time_sec);
906 }
907
908 void G1CollectorPolicy::record_concurrent_mark_cleanup_start() {
909 _mark_cleanup_start_sec = os::elapsedTime();
910 }
911
912 void G1CollectorPolicy::record_concurrent_mark_cleanup_completed() {
913 bool should_continue_with_reclaim = next_gc_should_be_mixed("request last young-only gc",
914 "skip last young-only gc");
915 collector_state()->set_last_young_gc(should_continue_with_reclaim);
916 // We skip the marking phase.
917 if (!should_continue_with_reclaim) {
918 abort_time_to_mixed_tracking();
919 }
920 collector_state()->set_in_marking_window(false);
921 }
922
923 void G1CollectorPolicy::record_concurrent_pause() {
924 if (_stop_world_start > 0.0) {
925 double yield_ms = (os::elapsedTime() - _stop_world_start) * 1000.0;
926 _trace_young_gen_time_data.record_yield_time(yield_ms);
927 }
928 }
929
930 double G1CollectorPolicy::average_time_ms(G1GCPhaseTimes::GCParPhases phase) const {
931 return phase_times()->average_time_ms(phase);
932 }
933
934 double G1CollectorPolicy::young_other_time_ms() const {
935 return phase_times()->young_cset_choice_time_ms() +
936 phase_times()->young_free_cset_time_ms();
937 }
938
939 double G1CollectorPolicy::non_young_other_time_ms() const {
940 return phase_times()->non_young_cset_choice_time_ms() +
941 phase_times()->non_young_free_cset_time_ms();
942
943 }
944
945 double G1CollectorPolicy::other_time_ms(double pause_time_ms) const {
946 return pause_time_ms -
947 average_time_ms(G1GCPhaseTimes::UpdateRS) -
948 average_time_ms(G1GCPhaseTimes::ScanRS) -
949 average_time_ms(G1GCPhaseTimes::ObjCopy) -
950 average_time_ms(G1GCPhaseTimes::Termination);
951 }
952
953 double G1CollectorPolicy::constant_other_time_ms(double pause_time_ms) const {
954 return other_time_ms(pause_time_ms) - young_other_time_ms() - non_young_other_time_ms();
955 }
956
957 bool G1CollectorPolicy::about_to_start_mixed_phase() const {
958 return _g1->concurrent_mark()->cmThread()->during_cycle() || collector_state()->last_young_gc();
959 }
960
961 bool G1CollectorPolicy::need_to_start_conc_mark(const char* source, size_t alloc_word_size) {
962 if (about_to_start_mixed_phase()) {
963 return false;
964 }
965
966 size_t marking_initiating_used_threshold = _ihop_control->get_conc_mark_start_threshold();
967
968 size_t cur_used_bytes = _g1->non_young_capacity_bytes();
969 size_t alloc_byte_size = alloc_word_size * HeapWordSize;
970 size_t marking_request_bytes = cur_used_bytes + alloc_byte_size;
971
972 bool result = false;
973 if (marking_request_bytes > marking_initiating_used_threshold) {
974 result = collector_state()->gcs_are_young() && !collector_state()->last_young_gc();
975 log_debug(gc, ergo, ihop)("%s occupancy: " SIZE_FORMAT "B allocation request: " SIZE_FORMAT "B threshold: " SIZE_FORMAT "B (%1.2f) source: %s",
976 result ? "Request concurrent cycle initiation (occupancy higher than threshold)" : "Do not request concurrent cycle initiation (still doing mixed collections)",
977 cur_used_bytes, alloc_byte_size, marking_initiating_used_threshold, (double) marking_initiating_used_threshold / _g1->capacity() * 100, source);
978 }
979
980 return result;
981 }
982
983 // Anything below that is considered to be zero
984 #define MIN_TIMER_GRANULARITY 0.0000001
985
986 void G1CollectorPolicy::record_collection_pause_end(double pause_time_ms, size_t cards_scanned) {
987 double end_time_sec = os::elapsedTime();
988
989 size_t cur_used_bytes = _g1->used();
990 assert(cur_used_bytes == _g1->recalculate_used(), "It should!");
991 bool last_pause_included_initial_mark = false;
992 bool update_stats = !_g1->evacuation_failed();
993
994 NOT_PRODUCT(_short_lived_surv_rate_group->print());
995
996 record_pause(young_gc_pause_kind(), end_time_sec - pause_time_ms / 1000.0, end_time_sec);
997
998 last_pause_included_initial_mark = collector_state()->during_initial_mark_pause();
999 if (last_pause_included_initial_mark) {
1000 record_concurrent_mark_init_end(0.0);
1001 } else {
1002 maybe_start_marking();
1003 }
1004
1005 double app_time_ms = (phase_times()->cur_collection_start_sec() * 1000.0 - _prev_collection_pause_end_ms);
1006 if (app_time_ms < MIN_TIMER_GRANULARITY) {
1007 // This usually happens due to the timer not having the required
1008 // granularity. Some Linuxes are the usual culprits.
1009 // We'll just set it to something (arbitrarily) small.
1010 app_time_ms = 1.0;
1011 }
1012
1013 if (update_stats) {
1014 _trace_young_gen_time_data.record_end_collection(pause_time_ms, phase_times());
1015 // We maintain the invariant that all objects allocated by mutator
1016 // threads will be allocated out of eden regions. So, we can use
1017 // the eden region number allocated since the previous GC to
1018 // calculate the application's allocate rate. The only exception
1019 // to that is humongous objects that are allocated separately. But
1020 // given that humongous object allocations do not really affect
1021 // either the pause's duration nor when the next pause will take
1022 // place we can safely ignore them here.
1023 uint regions_allocated = eden_cset_region_length();
1024 double alloc_rate_ms = (double) regions_allocated / app_time_ms;
1025 _alloc_rate_ms_seq->add(alloc_rate_ms);
1026
1027 double interval_ms =
1028 (end_time_sec - _recent_prev_end_times_for_all_gcs_sec->oldest()) * 1000.0;
1029 update_recent_gc_times(end_time_sec, pause_time_ms);
1030 _recent_avg_pause_time_ratio = _recent_gc_times_ms->sum()/interval_ms;
1031 if (recent_avg_pause_time_ratio() < 0.0 ||
1032 (recent_avg_pause_time_ratio() - 1.0 > 0.0)) {
1033 // Clip ratio between 0.0 and 1.0, and continue. This will be fixed in
1034 // CR 6902692 by redoing the manner in which the ratio is incrementally computed.
1035 if (_recent_avg_pause_time_ratio < 0.0) {
1036 _recent_avg_pause_time_ratio = 0.0;
1037 } else {
1038 assert(_recent_avg_pause_time_ratio - 1.0 > 0.0, "Ctl-point invariant");
1039 _recent_avg_pause_time_ratio = 1.0;
1040 }
1041 }
1042
1043 // Compute the ratio of just this last pause time to the entire time range stored
1044 // in the vectors. Comparing this pause to the entire range, rather than only the
1045 // most recent interval, has the effect of smoothing over a possible transient 'burst'
1046 // of more frequent pauses that don't really reflect a change in heap occupancy.
1047 // This reduces the likelihood of a needless heap expansion being triggered.
1048 _last_pause_time_ratio =
1049 (pause_time_ms * _recent_prev_end_times_for_all_gcs_sec->num()) / interval_ms;
1050 }
1051
1052 bool new_in_marking_window = collector_state()->in_marking_window();
1053 bool new_in_marking_window_im = false;
1054 if (last_pause_included_initial_mark) {
1055 new_in_marking_window = true;
1056 new_in_marking_window_im = true;
1057 }
1058
1059 if (collector_state()->last_young_gc()) {
1060 // This is supposed to to be the "last young GC" before we start
1061 // doing mixed GCs. Here we decide whether to start mixed GCs or not.
1062 assert(!last_pause_included_initial_mark, "The last young GC is not allowed to be an initial mark GC");
1063
1064 if (next_gc_should_be_mixed("start mixed GCs",
1065 "do not start mixed GCs")) {
1066 collector_state()->set_gcs_are_young(false);
1067 } else {
1068 // We aborted the mixed GC phase early.
1069 abort_time_to_mixed_tracking();
1070 }
1071
1072 collector_state()->set_last_young_gc(false);
1073 }
1074
1075 if (!collector_state()->last_gc_was_young()) {
1076 // This is a mixed GC. Here we decide whether to continue doing
1077 // mixed GCs or not.
1078 if (!next_gc_should_be_mixed("continue mixed GCs",
1079 "do not continue mixed GCs")) {
1080 collector_state()->set_gcs_are_young(true);
1081
1082 maybe_start_marking();
1083 }
1084 }
1085
1086 _short_lived_surv_rate_group->start_adding_regions();
1087 // Do that for any other surv rate groups
1088
1089 if (update_stats) {
1090 double cost_per_card_ms = 0.0;
1091 double cost_scan_hcc = average_time_ms(G1GCPhaseTimes::ScanHCC);
1092 if (_pending_cards > 0) {
1093 cost_per_card_ms = (average_time_ms(G1GCPhaseTimes::UpdateRS) - cost_scan_hcc) / (double) _pending_cards;
1094 _cost_per_card_ms_seq->add(cost_per_card_ms);
1095 }
1096 _cost_scan_hcc_seq->add(cost_scan_hcc);
1097
1098 double cost_per_entry_ms = 0.0;
1099 if (cards_scanned > 10) {
1100 cost_per_entry_ms = average_time_ms(G1GCPhaseTimes::ScanRS) / (double) cards_scanned;
1101 if (collector_state()->last_gc_was_young()) {
1102 _cost_per_entry_ms_seq->add(cost_per_entry_ms);
1103 } else {
1104 _mixed_cost_per_entry_ms_seq->add(cost_per_entry_ms);
1105 }
1106 }
1107
1108 if (_max_rs_lengths > 0) {
1109 double cards_per_entry_ratio =
1110 (double) cards_scanned / (double) _max_rs_lengths;
1111 if (collector_state()->last_gc_was_young()) {
1112 _young_cards_per_entry_ratio_seq->add(cards_per_entry_ratio);
1113 } else {
1114 _mixed_cards_per_entry_ratio_seq->add(cards_per_entry_ratio);
1115 }
1116 }
1117
1118 // This is defensive. For a while _max_rs_lengths could get
1119 // smaller than _recorded_rs_lengths which was causing
1120 // rs_length_diff to get very large and mess up the RSet length
1121 // predictions. The reason was unsafe concurrent updates to the
1122 // _inc_cset_recorded_rs_lengths field which the code below guards
1123 // against (see CR 7118202). This bug has now been fixed (see CR
1124 // 7119027). However, I'm still worried that
1125 // _inc_cset_recorded_rs_lengths might still end up somewhat
1126 // inaccurate. The concurrent refinement thread calculates an
1127 // RSet's length concurrently with other CR threads updating it
1128 // which might cause it to calculate the length incorrectly (if,
1129 // say, it's in mid-coarsening). So I'll leave in the defensive
1130 // conditional below just in case.
1131 size_t rs_length_diff = 0;
1132 if (_max_rs_lengths > _recorded_rs_lengths) {
1133 rs_length_diff = _max_rs_lengths - _recorded_rs_lengths;
1134 }
1135 _rs_length_diff_seq->add((double) rs_length_diff);
1136
1137 size_t freed_bytes = _heap_used_bytes_before_gc - cur_used_bytes;
1138 size_t copied_bytes = _collection_set_bytes_used_before - freed_bytes;
1139 double cost_per_byte_ms = 0.0;
1140
1141 if (copied_bytes > 0) {
1142 cost_per_byte_ms = average_time_ms(G1GCPhaseTimes::ObjCopy) / (double) copied_bytes;
1143 if (collector_state()->in_marking_window()) {
1144 _cost_per_byte_ms_during_cm_seq->add(cost_per_byte_ms);
1145 } else {
1146 _cost_per_byte_ms_seq->add(cost_per_byte_ms);
1147 }
1148 }
1149
1150 if (young_cset_region_length() > 0) {
1151 _young_other_cost_per_region_ms_seq->add(young_other_time_ms() /
1152 young_cset_region_length());
1153 }
1154
1155 if (old_cset_region_length() > 0) {
1156 _non_young_other_cost_per_region_ms_seq->add(non_young_other_time_ms() /
1157 old_cset_region_length());
1158 }
1159
1160 _constant_other_time_ms_seq->add(constant_other_time_ms(pause_time_ms));
1161
1162 _pending_cards_seq->add((double) _pending_cards);
1163 _rs_lengths_seq->add((double) _max_rs_lengths);
1164 }
1165
1166 collector_state()->set_in_marking_window(new_in_marking_window);
1167 collector_state()->set_in_marking_window_im(new_in_marking_window_im);
1168 _free_regions_at_end_of_collection = _g1->num_free_regions();
1169 // IHOP control wants to know the expected young gen length if it were not
1170 // restrained by the heap reserve. Using the actual length would make the
1171 // prediction too small and the limit the young gen every time we get to the
1172 // predicted target occupancy.
1173 size_t last_unrestrained_young_length = update_young_list_max_and_target_length();
1174 update_rs_lengths_prediction();
1175
1176 update_ihop_prediction(app_time_ms / 1000.0,
1177 _bytes_allocated_in_old_since_last_gc,
1178 last_unrestrained_young_length * HeapRegion::GrainBytes);
1179 _bytes_allocated_in_old_since_last_gc = 0;
1180
1181 _ihop_control->send_trace_event(_g1->gc_tracer_stw());
1182
1183 // Note that _mmu_tracker->max_gc_time() returns the time in seconds.
1184 double update_rs_time_goal_ms = _mmu_tracker->max_gc_time() * MILLIUNITS * G1RSetUpdatingPauseTimePercent / 100.0;
1185
1186 double scan_hcc_time_ms = average_time_ms(G1GCPhaseTimes::ScanHCC);
1187
1188 if (update_rs_time_goal_ms < scan_hcc_time_ms) {
1189 log_debug(gc, ergo, refine)("Adjust concurrent refinement thresholds (scanning the HCC expected to take longer than Update RS time goal)."
1190 "Update RS time goal: %1.2fms Scan HCC time: %1.2fms",
1191 update_rs_time_goal_ms, scan_hcc_time_ms);
1192
1193 update_rs_time_goal_ms = 0;
1194 } else {
1195 update_rs_time_goal_ms -= scan_hcc_time_ms;
1196 }
1197 adjust_concurrent_refinement(average_time_ms(G1GCPhaseTimes::UpdateRS) - scan_hcc_time_ms,
1198 phase_times()->sum_thread_work_items(G1GCPhaseTimes::UpdateRS),
1199 update_rs_time_goal_ms);
1200
1201 cset_chooser()->verify();
1202 }
1203
1204 G1IHOPControl* G1CollectorPolicy::create_ihop_control() const {
1205 if (G1UseAdaptiveIHOP) {
1206 return new G1AdaptiveIHOPControl(InitiatingHeapOccupancyPercent,
1207 G1CollectedHeap::heap()->max_capacity(),
1208 &_predictor,
1209 G1ReservePercent,
1210 G1HeapWastePercent);
1211 } else {
1212 return new G1StaticIHOPControl(InitiatingHeapOccupancyPercent,
1213 G1CollectedHeap::heap()->max_capacity());
1214 }
1215 }
1216
1217 void G1CollectorPolicy::update_ihop_prediction(double mutator_time_s,
1218 size_t mutator_alloc_bytes,
1219 size_t young_gen_size) {
1220 // Always try to update IHOP prediction. Even evacuation failures give information
1221 // about e.g. whether to start IHOP earlier next time.
1222
1223 // Avoid using really small application times that might create samples with
1224 // very high or very low values. They may be caused by e.g. back-to-back gcs.
1225 double const min_valid_time = 1e-6;
1226
1227 bool report = false;
1228
1229 double marking_to_mixed_time = -1.0;
1230 if (!collector_state()->last_gc_was_young() && _initial_mark_to_mixed.has_result()) {
1231 marking_to_mixed_time = _initial_mark_to_mixed.last_marking_time();
1232 assert(marking_to_mixed_time > 0.0,
1233 "Initial mark to mixed time must be larger than zero but is %.3f",
1234 marking_to_mixed_time);
1235 if (marking_to_mixed_time > min_valid_time) {
1236 _ihop_control->update_marking_length(marking_to_mixed_time);
1237 report = true;
1238 }
1239 }
1240
1241 // As an approximation for the young gc promotion rates during marking we use
1242 // all of them. In many applications there are only a few if any young gcs during
1243 // marking, which makes any prediction useless. This increases the accuracy of the
1244 // prediction.
1245 if (collector_state()->last_gc_was_young() && mutator_time_s > min_valid_time) {
1246 _ihop_control->update_allocation_info(mutator_time_s, mutator_alloc_bytes, young_gen_size);
1247 report = true;
1248 }
1249
1250 if (report) {
1251 report_ihop_statistics();
1252 }
1253 }
1254
1255 void G1CollectorPolicy::report_ihop_statistics() {
1256 _ihop_control->print();
1257 }
1258
1259 #define EXT_SIZE_FORMAT "%.1f%s"
1260 #define EXT_SIZE_PARAMS(bytes) \
1261 byte_size_in_proper_unit((double)(bytes)), \
1262 proper_unit_for_byte_size((bytes))
1263
1264 void G1CollectorPolicy::record_heap_size_info_at_start(bool full) {
1265 YoungList* young_list = _g1->young_list();
1266 _eden_used_bytes_before_gc = young_list->eden_used_bytes();
1267 _survivor_used_bytes_before_gc = young_list->survivor_used_bytes();
1268 _heap_capacity_bytes_before_gc = _g1->capacity();
1269 _old_used_bytes_before_gc = _g1->old_regions_count() * HeapRegion::GrainBytes;
1270 _humongous_used_bytes_before_gc = _g1->humongous_regions_count() * HeapRegion::GrainBytes;
1271 _heap_used_bytes_before_gc = _g1->used();
1272 _eden_capacity_bytes_before_gc = (_young_list_target_length * HeapRegion::GrainBytes) - _survivor_used_bytes_before_gc;
1273 _metaspace_used_bytes_before_gc = MetaspaceAux::used_bytes();
1274 }
1275
1276 void G1CollectorPolicy::print_detailed_heap_transition() const {
1277 YoungList* young_list = _g1->young_list();
1278
1279 size_t eden_used_bytes_after_gc = young_list->eden_used_bytes();
1280 size_t survivor_used_bytes_after_gc = young_list->survivor_used_bytes();
1281 size_t heap_used_bytes_after_gc = _g1->used();
1282 size_t old_used_bytes_after_gc = _g1->old_regions_count() * HeapRegion::GrainBytes;
1283 size_t humongous_used_bytes_after_gc = _g1->humongous_regions_count() * HeapRegion::GrainBytes;
1284
1285 size_t heap_capacity_bytes_after_gc = _g1->capacity();
1286 size_t eden_capacity_bytes_after_gc =
1287 (_young_list_target_length * HeapRegion::GrainBytes) - survivor_used_bytes_after_gc;
1288 size_t survivor_capacity_bytes_after_gc = _max_survivor_regions * HeapRegion::GrainBytes;
1289
1290 log_info(gc, heap)("Eden: " SIZE_FORMAT "K->" SIZE_FORMAT "K(" SIZE_FORMAT "K)",
1291 _eden_used_bytes_before_gc / K, eden_used_bytes_after_gc /K, eden_capacity_bytes_after_gc /K);
1292 log_info(gc, heap)("Survivor: " SIZE_FORMAT "K->" SIZE_FORMAT "K(" SIZE_FORMAT "K)",
1293 _survivor_used_bytes_before_gc / K, survivor_used_bytes_after_gc /K, survivor_capacity_bytes_after_gc /K);
1294 log_info(gc, heap)("Old: " SIZE_FORMAT "K->" SIZE_FORMAT "K",
1295 _old_used_bytes_before_gc / K, old_used_bytes_after_gc /K);
1296 log_info(gc, heap)("Humongous: " SIZE_FORMAT "K->" SIZE_FORMAT "K",
1297 _humongous_used_bytes_before_gc / K, humongous_used_bytes_after_gc /K);
1298
1299 MetaspaceAux::print_metaspace_change(_metaspace_used_bytes_before_gc);
1300 }
1301
1302 void G1CollectorPolicy::print_phases(double pause_time_sec) {
1303 phase_times()->print(pause_time_sec);
1304 }
1305
1306 void G1CollectorPolicy::adjust_concurrent_refinement(double update_rs_time,
1307 double update_rs_processed_buffers,
1308 double goal_ms) {
1309 DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
1310 ConcurrentG1Refine *cg1r = G1CollectedHeap::heap()->concurrent_g1_refine();
1311
1312 if (G1UseAdaptiveConcRefinement) {
1313 const int k_gy = 3, k_gr = 6;
1314 const double inc_k = 1.1, dec_k = 0.9;
1315
1316 int g = cg1r->green_zone();
1317 if (update_rs_time > goal_ms) {
1318 g = (int)(g * dec_k); // Can become 0, that's OK. That would mean a mutator-only processing.
1319 } else {
1320 if (update_rs_time < goal_ms && update_rs_processed_buffers > g) {
1321 g = (int)MAX2(g * inc_k, g + 1.0);
1322 }
1323 }
1324 // Change the refinement threads params
1325 cg1r->set_green_zone(g);
1326 cg1r->set_yellow_zone(g * k_gy);
1327 cg1r->set_red_zone(g * k_gr);
1328 cg1r->reinitialize_threads();
1329
1330 int processing_threshold_delta = MAX2((int)(cg1r->green_zone() * _predictor.sigma()), 1);
1331 int processing_threshold = MIN2(cg1r->green_zone() + processing_threshold_delta,
1332 cg1r->yellow_zone());
1333 // Change the barrier params
1334 dcqs.set_process_completed_threshold(processing_threshold);
1335 dcqs.set_max_completed_queue(cg1r->red_zone());
1336 }
1337
1338 int curr_queue_size = dcqs.completed_buffers_num();
1339 if (curr_queue_size >= cg1r->yellow_zone()) {
1340 dcqs.set_completed_queue_padding(curr_queue_size);
1341 } else {
1342 dcqs.set_completed_queue_padding(0);
1343 }
1344 dcqs.notify_if_necessary();
1345 }
1346
1347 size_t G1CollectorPolicy::predict_rs_length_diff() const {
1348 return (size_t) get_new_prediction(_rs_length_diff_seq);
1349 }
1350
1351 double G1CollectorPolicy::predict_alloc_rate_ms() const {
1352 return get_new_prediction(_alloc_rate_ms_seq);
1353 }
1354
1355 double G1CollectorPolicy::predict_cost_per_card_ms() const {
1356 return get_new_prediction(_cost_per_card_ms_seq);
1357 }
1358
1359 double G1CollectorPolicy::predict_scan_hcc_ms() const {
1360 return get_new_prediction(_cost_scan_hcc_seq);
1361 }
1362
1363 double G1CollectorPolicy::predict_rs_update_time_ms(size_t pending_cards) const {
1364 return pending_cards * predict_cost_per_card_ms() + predict_scan_hcc_ms();
1365 }
1366
1367 double G1CollectorPolicy::predict_young_cards_per_entry_ratio() const {
1368 return get_new_prediction(_young_cards_per_entry_ratio_seq);
1369 }
1370
1371 double G1CollectorPolicy::predict_mixed_cards_per_entry_ratio() const {
1372 if (_mixed_cards_per_entry_ratio_seq->num() < 2) {
1373 return predict_young_cards_per_entry_ratio();
1374 } else {
1375 return get_new_prediction(_mixed_cards_per_entry_ratio_seq);
1376 }
1377 }
1378
1379 size_t G1CollectorPolicy::predict_young_card_num(size_t rs_length) const {
1380 return (size_t) (rs_length * predict_young_cards_per_entry_ratio());
1381 }
1382
1383 size_t G1CollectorPolicy::predict_non_young_card_num(size_t rs_length) const {
1384 return (size_t)(rs_length * predict_mixed_cards_per_entry_ratio());
1385 }
1386
1387 double G1CollectorPolicy::predict_rs_scan_time_ms(size_t card_num) const {
1388 if (collector_state()->gcs_are_young()) {
1389 return card_num * get_new_prediction(_cost_per_entry_ms_seq);
1390 } else {
1391 return predict_mixed_rs_scan_time_ms(card_num);
1392 }
1393 }
1394
1395 double G1CollectorPolicy::predict_mixed_rs_scan_time_ms(size_t card_num) const {
1396 if (_mixed_cost_per_entry_ms_seq->num() < 3) {
1397 return card_num * get_new_prediction(_cost_per_entry_ms_seq);
1398 } else {
1399 return card_num * get_new_prediction(_mixed_cost_per_entry_ms_seq);
1400 }
1401 }
1402
1403 double G1CollectorPolicy::predict_object_copy_time_ms_during_cm(size_t bytes_to_copy) const {
1404 if (_cost_per_byte_ms_during_cm_seq->num() < 3) {
1405 return (1.1 * bytes_to_copy) * get_new_prediction(_cost_per_byte_ms_seq);
1406 } else {
1407 return bytes_to_copy * get_new_prediction(_cost_per_byte_ms_during_cm_seq);
1408 }
1409 }
1410
1411 double G1CollectorPolicy::predict_object_copy_time_ms(size_t bytes_to_copy) const {
1412 if (collector_state()->during_concurrent_mark()) {
1413 return predict_object_copy_time_ms_during_cm(bytes_to_copy);
1414 } else {
1415 return bytes_to_copy * get_new_prediction(_cost_per_byte_ms_seq);
1416 }
1417 }
1418
1419 double G1CollectorPolicy::predict_constant_other_time_ms() const {
1420 return get_new_prediction(_constant_other_time_ms_seq);
1421 }
1422
1423 double G1CollectorPolicy::predict_young_other_time_ms(size_t young_num) const {
1424 return young_num * get_new_prediction(_young_other_cost_per_region_ms_seq);
1425 }
1426
1427 double G1CollectorPolicy::predict_non_young_other_time_ms(size_t non_young_num) const {
1428 return non_young_num * get_new_prediction(_non_young_other_cost_per_region_ms_seq);
1429 }
1430
1431 double G1CollectorPolicy::predict_remark_time_ms() const {
1432 return get_new_prediction(_concurrent_mark_remark_times_ms);
1433 }
1434
1435 double G1CollectorPolicy::predict_cleanup_time_ms() const {
1436 return get_new_prediction(_concurrent_mark_cleanup_times_ms);
1437 }
1438
1439 double G1CollectorPolicy::predict_yg_surv_rate(int age, SurvRateGroup* surv_rate_group) const {
1440 TruncatedSeq* seq = surv_rate_group->get_seq(age);
1441 guarantee(seq->num() > 0, "There should be some young gen survivor samples available. Tried to access with age %d", age);
1442 double pred = get_new_prediction(seq);
1443 if (pred > 1.0) {
1444 pred = 1.0;
1445 }
1446 return pred;
1447 }
1448
1449 double G1CollectorPolicy::predict_yg_surv_rate(int age) const {
1450 return predict_yg_surv_rate(age, _short_lived_surv_rate_group);
1451 }
1452
1453 double G1CollectorPolicy::accum_yg_surv_rate_pred(int age) const {
1454 return _short_lived_surv_rate_group->accum_surv_rate_pred(age);
1455 }
1456
1457 double G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards,
1458 size_t scanned_cards) const {
1459 return
1460 predict_rs_update_time_ms(pending_cards) +
1461 predict_rs_scan_time_ms(scanned_cards) +
1462 predict_constant_other_time_ms();
1463 }
1464
1465 double G1CollectorPolicy::predict_base_elapsed_time_ms(size_t pending_cards) const {
1466 size_t rs_length = predict_rs_length_diff();
1467 size_t card_num;
1468 if (collector_state()->gcs_are_young()) {
1469 card_num = predict_young_card_num(rs_length);
1470 } else {
1471 card_num = predict_non_young_card_num(rs_length);
1472 }
1473 return predict_base_elapsed_time_ms(pending_cards, card_num);
1474 }
1475
1476 size_t G1CollectorPolicy::predict_bytes_to_copy(HeapRegion* hr) const {
1477 size_t bytes_to_copy;
1478 if (hr->is_marked())
1479 bytes_to_copy = hr->max_live_bytes();
1480 else {
1481 assert(hr->is_young() && hr->age_in_surv_rate_group() != -1, "invariant");
1482 int age = hr->age_in_surv_rate_group();
1483 double yg_surv_rate = predict_yg_surv_rate(age, hr->surv_rate_group());
1484 bytes_to_copy = (size_t) (hr->used() * yg_surv_rate);
1485 }
1486 return bytes_to_copy;
1487 }
1488
1489 double G1CollectorPolicy::predict_region_elapsed_time_ms(HeapRegion* hr,
1490 bool for_young_gc) const {
1491 size_t rs_length = hr->rem_set()->occupied();
1492 size_t card_num;
1493
1494 // Predicting the number of cards is based on which type of GC
1495 // we're predicting for.
1496 if (for_young_gc) {
1497 card_num = predict_young_card_num(rs_length);
1498 } else {
1499 card_num = predict_non_young_card_num(rs_length);
1500 }
1501 size_t bytes_to_copy = predict_bytes_to_copy(hr);
1502
1503 double region_elapsed_time_ms =
1504 predict_rs_scan_time_ms(card_num) +
1505 predict_object_copy_time_ms(bytes_to_copy);
1506
1507 // The prediction of the "other" time for this region is based
1508 // upon the region type and NOT the GC type.
1509 if (hr->is_young()) {
1510 region_elapsed_time_ms += predict_young_other_time_ms(1);
1511 } else {
1512 region_elapsed_time_ms += predict_non_young_other_time_ms(1);
1513 }
1514 return region_elapsed_time_ms;
1515 }
1516
1517 void G1CollectorPolicy::init_cset_region_lengths(uint eden_cset_region_length,
1518 uint survivor_cset_region_length) {
1519 _eden_cset_region_length = eden_cset_region_length;
1520 _survivor_cset_region_length = survivor_cset_region_length;
1521 _old_cset_region_length = 0;
1522 }
1523
1524 void G1CollectorPolicy::set_recorded_rs_lengths(size_t rs_lengths) {
1525 _recorded_rs_lengths = rs_lengths;
1526 }
1527
1528 void G1CollectorPolicy::update_recent_gc_times(double end_time_sec,
1529 double elapsed_ms) {
1530 _recent_gc_times_ms->add(elapsed_ms);
1531 _recent_prev_end_times_for_all_gcs_sec->add(end_time_sec);
1532 _prev_collection_pause_end_ms = end_time_sec * 1000.0;
1533 }
1534
1535 void G1CollectorPolicy::clear_ratio_check_data() {
1536 _ratio_over_threshold_count = 0;
1537 _ratio_over_threshold_sum = 0.0;
1538 _pauses_since_start = 0;
1539 }
1540
1541 size_t G1CollectorPolicy::expansion_amount() {
1542 double recent_gc_overhead = recent_avg_pause_time_ratio() * 100.0;
1543 double last_gc_overhead = _last_pause_time_ratio * 100.0;
1544 double threshold = _gc_overhead_perc;
1545 size_t expand_bytes = 0;
1546
1547 // If the heap is at less than half its maximum size, scale the threshold down,
1548 // to a limit of 1. Thus the smaller the heap is, the more likely it is to expand,
1549 // though the scaling code will likely keep the increase small.
1550 if (_g1->capacity() <= _g1->max_capacity() / 2) {
1551 threshold *= (double)_g1->capacity() / (double)(_g1->max_capacity() / 2);
1552 threshold = MAX2(threshold, 1.0);
1553 }
1554
1555 // If the last GC time ratio is over the threshold, increment the count of
1556 // times it has been exceeded, and add this ratio to the sum of exceeded
1557 // ratios.
1558 if (last_gc_overhead > threshold) {
1559 _ratio_over_threshold_count++;
1560 _ratio_over_threshold_sum += last_gc_overhead;
1561 }
1562
1563 // Check if we've had enough GC time ratio checks that were over the
1564 // threshold to trigger an expansion. We'll also expand if we've
1565 // reached the end of the history buffer and the average of all entries
1566 // is still over the threshold. This indicates a smaller number of GCs were
1567 // long enough to make the average exceed the threshold.
1568 bool filled_history_buffer = _pauses_since_start == NumPrevPausesForHeuristics;
1569 if ((_ratio_over_threshold_count == MinOverThresholdForGrowth) ||
1570 (filled_history_buffer && (recent_gc_overhead > threshold))) {
1571 size_t min_expand_bytes = HeapRegion::GrainBytes;
1572 size_t reserved_bytes = _g1->max_capacity();
1573 size_t committed_bytes = _g1->capacity();
1574 size_t uncommitted_bytes = reserved_bytes - committed_bytes;
1575 size_t expand_bytes_via_pct =
1576 uncommitted_bytes * G1ExpandByPercentOfAvailable / 100;
1577 double scale_factor = 1.0;
1578
1579 // If the current size is less than 1/4 of the Initial heap size, expand
1580 // by half of the delta between the current and Initial sizes. IE, grow
1581 // back quickly.
1582 //
1583 // Otherwise, take the current size, or G1ExpandByPercentOfAvailable % of
1584 // the available expansion space, whichever is smaller, as the base
1585 // expansion size. Then possibly scale this size according to how much the
1586 // threshold has (on average) been exceeded by. If the delta is small
1587 // (less than the StartScaleDownAt value), scale the size down linearly, but
1588 // not by less than MinScaleDownFactor. If the delta is large (greater than
1589 // the StartScaleUpAt value), scale up, but adding no more than MaxScaleUpFactor
1590 // times the base size. The scaling will be linear in the range from
1591 // StartScaleUpAt to (StartScaleUpAt + ScaleUpRange). In other words,
1592 // ScaleUpRange sets the rate of scaling up.
1593 if (committed_bytes < InitialHeapSize / 4) {
1594 expand_bytes = (InitialHeapSize - committed_bytes) / 2;
1595 } else {
1596 double const MinScaleDownFactor = 0.2;
1597 double const MaxScaleUpFactor = 2;
1598 double const StartScaleDownAt = _gc_overhead_perc;
1599 double const StartScaleUpAt = _gc_overhead_perc * 1.5;
1600 double const ScaleUpRange = _gc_overhead_perc * 2.0;
1601
1602 double ratio_delta;
1603 if (filled_history_buffer) {
1604 ratio_delta = recent_gc_overhead - threshold;
1605 } else {
1606 ratio_delta = (_ratio_over_threshold_sum/_ratio_over_threshold_count) - threshold;
1607 }
1608
1609 expand_bytes = MIN2(expand_bytes_via_pct, committed_bytes);
1610 if (ratio_delta < StartScaleDownAt) {
1611 scale_factor = ratio_delta / StartScaleDownAt;
1612 scale_factor = MAX2(scale_factor, MinScaleDownFactor);
1613 } else if (ratio_delta > StartScaleUpAt) {
1614 scale_factor = 1 + ((ratio_delta - StartScaleUpAt) / ScaleUpRange);
1615 scale_factor = MIN2(scale_factor, MaxScaleUpFactor);
1616 }
1617 }
1618
1619 log_debug(gc, ergo, heap)("Attempt heap expansion (recent GC overhead higher than threshold after GC) "
1620 "recent GC overhead: %1.2f %% threshold: %1.2f %% uncommitted: " SIZE_FORMAT "B base expansion amount and scale: " SIZE_FORMAT "B (%1.2f%%)",
1621 recent_gc_overhead, threshold, uncommitted_bytes, expand_bytes, scale_factor * 100);
1622
1623 expand_bytes = static_cast<size_t>(expand_bytes * scale_factor);
1624
1625 // Ensure the expansion size is at least the minimum growth amount
1626 // and at most the remaining uncommitted byte size.
1627 expand_bytes = MAX2(expand_bytes, min_expand_bytes);
1628 expand_bytes = MIN2(expand_bytes, uncommitted_bytes);
1629
1630 clear_ratio_check_data();
1631 } else {
1632 // An expansion was not triggered. If we've started counting, increment
1633 // the number of checks we've made in the current window. If we've
1634 // reached the end of the window without resizing, clear the counters to
1635 // start again the next time we see a ratio above the threshold.
1636 if (_ratio_over_threshold_count > 0) {
1637 _pauses_since_start++;
1638 if (_pauses_since_start > NumPrevPausesForHeuristics) {
1639 clear_ratio_check_data();
1640 }
1641 }
1642 }
1643
1644 return expand_bytes;
1645 }
1646
1647 void G1CollectorPolicy::print_tracing_info() const {
1648 _trace_young_gen_time_data.print();
1649 _trace_old_gen_time_data.print();
1650 }
1651
1652 void G1CollectorPolicy::print_yg_surv_rate_info() const {
1653 #ifndef PRODUCT
1654 _short_lived_surv_rate_group->print_surv_rate_summary();
1655 // add this call for any other surv rate groups
1656 #endif // PRODUCT
1657 }
1658
1659 bool G1CollectorPolicy::is_young_list_full() const {
1660 uint young_list_length = _g1->young_list()->length();
1661 uint young_list_target_length = _young_list_target_length;
1662 return young_list_length >= young_list_target_length;
1663 }
1664
1665 bool G1CollectorPolicy::can_expand_young_list() const {
1666 uint young_list_length = _g1->young_list()->length();
1667 uint young_list_max_length = _young_list_max_length;
1668 return young_list_length < young_list_max_length;
1669 }
1670
1671 void G1CollectorPolicy::update_max_gc_locker_expansion() {
1672 uint expansion_region_num = 0;
1673 if (GCLockerEdenExpansionPercent > 0) {
1674 double perc = (double) GCLockerEdenExpansionPercent / 100.0;
1675 double expansion_region_num_d = perc * (double) _young_list_target_length;
1676 // We use ceiling so that if expansion_region_num_d is > 0.0 (but
1677 // less than 1.0) we'll get 1.
1678 expansion_region_num = (uint) ceil(expansion_region_num_d);
1679 } else {
1680 assert(expansion_region_num == 0, "sanity");
1681 }
1682 _young_list_max_length = _young_list_target_length + expansion_region_num;
1683 assert(_young_list_target_length <= _young_list_max_length, "post-condition");
1684 }
1685
1686 // Calculates survivor space parameters.
1687 void G1CollectorPolicy::update_survivors_policy() {
1688 double max_survivor_regions_d =
1689 (double) _young_list_target_length / (double) SurvivorRatio;
1690 // We use ceiling so that if max_survivor_regions_d is > 0.0 (but
1691 // smaller than 1.0) we'll get 1.
1692 _max_survivor_regions = (uint) ceil(max_survivor_regions_d);
1693
1694 _tenuring_threshold = _survivors_age_table.compute_tenuring_threshold(
1695 HeapRegion::GrainWords * _max_survivor_regions, counters());
1696 }
1697
1698 bool G1CollectorPolicy::force_initial_mark_if_outside_cycle(GCCause::Cause gc_cause) {
1699 // We actually check whether we are marking here and not if we are in a
1700 // reclamation phase. This means that we will schedule a concurrent mark
1701 // even while we are still in the process of reclaiming memory.
1702 bool during_cycle = _g1->concurrent_mark()->cmThread()->during_cycle();
1703 if (!during_cycle) {
1704 log_debug(gc, ergo)("Request concurrent cycle initiation (requested by GC cause). GC cause: %s", GCCause::to_string(gc_cause));
1705 collector_state()->set_initiate_conc_mark_if_possible(true);
1706 return true;
1707 } else {
1708 log_debug(gc, ergo)("Do not request concurrent cycle initiation (concurrent cycle already in progress). GC cause: %s", GCCause::to_string(gc_cause));
1709 return false;
1710 }
1711 }
1712
1713 void G1CollectorPolicy::initiate_conc_mark() {
1714 collector_state()->set_during_initial_mark_pause(true);
1715 collector_state()->set_initiate_conc_mark_if_possible(false);
1716 }
1717
1718 void G1CollectorPolicy::decide_on_conc_mark_initiation() {
1719 // We are about to decide on whether this pause will be an
1720 // initial-mark pause.
1721
1722 // First, collector_state()->during_initial_mark_pause() should not be already set. We
1723 // will set it here if we have to. However, it should be cleared by
1724 // the end of the pause (it's only set for the duration of an
1725 // initial-mark pause).
1726 assert(!collector_state()->during_initial_mark_pause(), "pre-condition");
1727
1728 if (collector_state()->initiate_conc_mark_if_possible()) {
1729 // We had noticed on a previous pause that the heap occupancy has
1730 // gone over the initiating threshold and we should start a
1731 // concurrent marking cycle. So we might initiate one.
1732
1733 if (!about_to_start_mixed_phase() && collector_state()->gcs_are_young()) {
1734 // Initiate a new initial mark if there is no marking or reclamation going on.
1735 initiate_conc_mark();
1736 log_debug(gc, ergo)("Initiate concurrent cycle (concurrent cycle initiation requested)");
1737 } else if (_g1->is_user_requested_concurrent_full_gc(_g1->gc_cause())) {
1738 // Initiate a user requested initial mark. An initial mark must be young only
1739 // GC, so the collector state must be updated to reflect this.
1740 collector_state()->set_gcs_are_young(true);
1741 collector_state()->set_last_young_gc(false);
1742
1743 abort_time_to_mixed_tracking();
1744 initiate_conc_mark();
1745 log_debug(gc, ergo)("Initiate concurrent cycle (user requested concurrent cycle)");
1746 } else {
1747 // The concurrent marking thread is still finishing up the
1748 // previous cycle. If we start one right now the two cycles
1749 // overlap. In particular, the concurrent marking thread might
1750 // be in the process of clearing the next marking bitmap (which
1751 // we will use for the next cycle if we start one). Starting a
1752 // cycle now will be bad given that parts of the marking
1753 // information might get cleared by the marking thread. And we
1754 // cannot wait for the marking thread to finish the cycle as it
1755 // periodically yields while clearing the next marking bitmap
1756 // and, if it's in a yield point, it's waiting for us to
1757 // finish. So, at this point we will not start a cycle and we'll
1758 // let the concurrent marking thread complete the last one.
1759 log_debug(gc, ergo)("Do not initiate concurrent cycle (concurrent cycle already in progress)");
1760 }
1761 }
1762 }
1763
1764 class ParKnownGarbageHRClosure: public HeapRegionClosure {
1765 G1CollectedHeap* _g1h;
1766 CSetChooserParUpdater _cset_updater;
1767
1768 public:
1769 ParKnownGarbageHRClosure(CollectionSetChooser* hrSorted,
1770 uint chunk_size) :
1771 _g1h(G1CollectedHeap::heap()),
1772 _cset_updater(hrSorted, true /* parallel */, chunk_size) { }
1773
1774 bool doHeapRegion(HeapRegion* r) {
1775 // Do we have any marking information for this region?
1776 if (r->is_marked()) {
1777 // We will skip any region that's currently used as an old GC
1778 // alloc region (we should not consider those for collection
1779 // before we fill them up).
1780 if (_cset_updater.should_add(r) && !_g1h->is_old_gc_alloc_region(r)) {
1781 _cset_updater.add_region(r);
1782 }
1783 }
1784 return false;
1785 }
1786 };
1787
1788 class ParKnownGarbageTask: public AbstractGangTask {
1789 CollectionSetChooser* _hrSorted;
1790 uint _chunk_size;
1791 G1CollectedHeap* _g1;
1792 HeapRegionClaimer _hrclaimer;
1793
1794 public:
1795 ParKnownGarbageTask(CollectionSetChooser* hrSorted, uint chunk_size, uint n_workers) :
1796 AbstractGangTask("ParKnownGarbageTask"),
1797 _hrSorted(hrSorted), _chunk_size(chunk_size),
1798 _g1(G1CollectedHeap::heap()), _hrclaimer(n_workers) {}
1799
1800 void work(uint worker_id) {
1801 ParKnownGarbageHRClosure parKnownGarbageCl(_hrSorted, _chunk_size);
1802 _g1->heap_region_par_iterate(&parKnownGarbageCl, worker_id, &_hrclaimer);
1803 }
1804 };
1805
1806 uint G1CollectorPolicy::calculate_parallel_work_chunk_size(uint n_workers, uint n_regions) const {
1807 assert(n_workers > 0, "Active gc workers should be greater than 0");
1808 const uint overpartition_factor = 4;
1809 const uint min_chunk_size = MAX2(n_regions / n_workers, 1U);
1810 return MAX2(n_regions / (n_workers * overpartition_factor), min_chunk_size);
1811 }
1812
1813 void G1CollectorPolicy::record_concurrent_mark_cleanup_end() {
1814 cset_chooser()->clear();
1815
1816 WorkGang* workers = _g1->workers();
1817 uint n_workers = workers->active_workers();
1818
1819 uint n_regions = _g1->num_regions();
1820 uint chunk_size = calculate_parallel_work_chunk_size(n_workers, n_regions);
1821 cset_chooser()->prepare_for_par_region_addition(n_workers, n_regions, chunk_size);
1822 ParKnownGarbageTask par_known_garbage_task(cset_chooser(), chunk_size, n_workers);
1823 workers->run_task(&par_known_garbage_task);
1824
1825 cset_chooser()->sort_regions();
1826
1827 double end_sec = os::elapsedTime();
1828 double elapsed_time_ms = (end_sec - _mark_cleanup_start_sec) * 1000.0;
1829 _concurrent_mark_cleanup_times_ms->add(elapsed_time_ms);
1830 _prev_collection_pause_end_ms += elapsed_time_ms;
1831
1832 record_pause(Cleanup, _mark_cleanup_start_sec, end_sec);
1833 }
1834
1835 // Add the heap region at the head of the non-incremental collection set
1836 void G1CollectorPolicy::add_old_region_to_cset(HeapRegion* hr) {
1837 assert(_inc_cset_build_state == Active, "Precondition");
1838 assert(hr->is_old(), "the region should be old");
1839
1840 assert(!hr->in_collection_set(), "should not already be in the CSet");
1841 _g1->register_old_region_with_cset(hr);
1842 hr->set_next_in_collection_set(_collection_set);
1843 _collection_set = hr;
1844 _collection_set_bytes_used_before += hr->used();
1845 size_t rs_length = hr->rem_set()->occupied();
1846 _recorded_rs_lengths += rs_length;
1847 _old_cset_region_length += 1;
1848 }
1849
1850 // Initialize the per-collection-set information
1851 void G1CollectorPolicy::start_incremental_cset_building() {
1852 assert(_inc_cset_build_state == Inactive, "Precondition");
1853
1854 _inc_cset_head = NULL;
1855 _inc_cset_tail = NULL;
1856 _inc_cset_bytes_used_before = 0;
1857
1858 _inc_cset_max_finger = 0;
1859 _inc_cset_recorded_rs_lengths = 0;
1860 _inc_cset_recorded_rs_lengths_diffs = 0;
1861 _inc_cset_predicted_elapsed_time_ms = 0.0;
1862 _inc_cset_predicted_elapsed_time_ms_diffs = 0.0;
1863 _inc_cset_build_state = Active;
1864 }
1865
1866 void G1CollectorPolicy::finalize_incremental_cset_building() {
1867 assert(_inc_cset_build_state == Active, "Precondition");
1868 assert(SafepointSynchronize::is_at_safepoint(), "should be at a safepoint");
1869
1870 // The two "main" fields, _inc_cset_recorded_rs_lengths and
1871 // _inc_cset_predicted_elapsed_time_ms, are updated by the thread
1872 // that adds a new region to the CSet. Further updates by the
1873 // concurrent refinement thread that samples the young RSet lengths
1874 // are accumulated in the *_diffs fields. Here we add the diffs to
1875 // the "main" fields.
1876
1877 if (_inc_cset_recorded_rs_lengths_diffs >= 0) {
1878 _inc_cset_recorded_rs_lengths += _inc_cset_recorded_rs_lengths_diffs;
1879 } else {
1880 // This is defensive. The diff should in theory be always positive
1881 // as RSets can only grow between GCs. However, given that we
1882 // sample their size concurrently with other threads updating them
1883 // it's possible that we might get the wrong size back, which
1884 // could make the calculations somewhat inaccurate.
1885 size_t diffs = (size_t) (-_inc_cset_recorded_rs_lengths_diffs);
1886 if (_inc_cset_recorded_rs_lengths >= diffs) {
1887 _inc_cset_recorded_rs_lengths -= diffs;
1888 } else {
1889 _inc_cset_recorded_rs_lengths = 0;
1890 }
1891 }
1892 _inc_cset_predicted_elapsed_time_ms +=
1893 _inc_cset_predicted_elapsed_time_ms_diffs;
1894
1895 _inc_cset_recorded_rs_lengths_diffs = 0;
1896 _inc_cset_predicted_elapsed_time_ms_diffs = 0.0;
1897 }
1898
1899 void G1CollectorPolicy::add_to_incremental_cset_info(HeapRegion* hr, size_t rs_length) {
1900 // This routine is used when:
1901 // * adding survivor regions to the incremental cset at the end of an
1902 // evacuation pause,
1903 // * adding the current allocation region to the incremental cset
1904 // when it is retired, and
1905 // * updating existing policy information for a region in the
1906 // incremental cset via young list RSet sampling.
1907 // Therefore this routine may be called at a safepoint by the
1908 // VM thread, or in-between safepoints by mutator threads (when
1909 // retiring the current allocation region) or a concurrent
1910 // refine thread (RSet sampling).
1911
1912 double region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
1913 size_t used_bytes = hr->used();
1914 _inc_cset_recorded_rs_lengths += rs_length;
1915 _inc_cset_predicted_elapsed_time_ms += region_elapsed_time_ms;
1916 _inc_cset_bytes_used_before += used_bytes;
1917
1918 // Cache the values we have added to the aggregated information
1919 // in the heap region in case we have to remove this region from
1920 // the incremental collection set, or it is updated by the
1921 // rset sampling code
1922 hr->set_recorded_rs_length(rs_length);
1923 hr->set_predicted_elapsed_time_ms(region_elapsed_time_ms);
1924 }
1925
1926 void G1CollectorPolicy::update_incremental_cset_info(HeapRegion* hr,
1927 size_t new_rs_length) {
1928 // Update the CSet information that is dependent on the new RS length
1929 assert(hr->is_young(), "Precondition");
1930 assert(!SafepointSynchronize::is_at_safepoint(),
1931 "should not be at a safepoint");
1932
1933 // We could have updated _inc_cset_recorded_rs_lengths and
1934 // _inc_cset_predicted_elapsed_time_ms directly but we'd need to do
1935 // that atomically, as this code is executed by a concurrent
1936 // refinement thread, potentially concurrently with a mutator thread
1937 // allocating a new region and also updating the same fields. To
1938 // avoid the atomic operations we accumulate these updates on two
1939 // separate fields (*_diffs) and we'll just add them to the "main"
1940 // fields at the start of a GC.
1941
1942 ssize_t old_rs_length = (ssize_t) hr->recorded_rs_length();
1943 ssize_t rs_lengths_diff = (ssize_t) new_rs_length - old_rs_length;
1944 _inc_cset_recorded_rs_lengths_diffs += rs_lengths_diff;
1945
1946 double old_elapsed_time_ms = hr->predicted_elapsed_time_ms();
1947 double new_region_elapsed_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
1948 double elapsed_ms_diff = new_region_elapsed_time_ms - old_elapsed_time_ms;
1949 _inc_cset_predicted_elapsed_time_ms_diffs += elapsed_ms_diff;
1950
1951 hr->set_recorded_rs_length(new_rs_length);
1952 hr->set_predicted_elapsed_time_ms(new_region_elapsed_time_ms);
1953 }
1954
1955 void G1CollectorPolicy::add_region_to_incremental_cset_common(HeapRegion* hr) {
1956 assert(hr->is_young(), "invariant");
1957 assert(hr->young_index_in_cset() > -1, "should have already been set");
1958 assert(_inc_cset_build_state == Active, "Precondition");
1959
1960 // We need to clear and set the cached recorded/cached collection set
1961 // information in the heap region here (before the region gets added
1962 // to the collection set). An individual heap region's cached values
1963 // are calculated, aggregated with the policy collection set info,
1964 // and cached in the heap region here (initially) and (subsequently)
1965 // by the Young List sampling code.
1966
1967 size_t rs_length = hr->rem_set()->occupied();
1968 add_to_incremental_cset_info(hr, rs_length);
1969
1970 HeapWord* hr_end = hr->end();
1971 _inc_cset_max_finger = MAX2(_inc_cset_max_finger, hr_end);
1972
1973 assert(!hr->in_collection_set(), "invariant");
1974 _g1->register_young_region_with_cset(hr);
1975 assert(hr->next_in_collection_set() == NULL, "invariant");
1976 }
1977
1978 // Add the region at the RHS of the incremental cset
1979 void G1CollectorPolicy::add_region_to_incremental_cset_rhs(HeapRegion* hr) {
1980 // We should only ever be appending survivors at the end of a pause
1981 assert(hr->is_survivor(), "Logic");
1982
1983 // Do the 'common' stuff
1984 add_region_to_incremental_cset_common(hr);
1985
1986 // Now add the region at the right hand side
1987 if (_inc_cset_tail == NULL) {
1988 assert(_inc_cset_head == NULL, "invariant");
1989 _inc_cset_head = hr;
1990 } else {
1991 _inc_cset_tail->set_next_in_collection_set(hr);
1992 }
1993 _inc_cset_tail = hr;
1994 }
1995
1996 // Add the region to the LHS of the incremental cset
1997 void G1CollectorPolicy::add_region_to_incremental_cset_lhs(HeapRegion* hr) {
1998 // Survivors should be added to the RHS at the end of a pause
1999 assert(hr->is_eden(), "Logic");
2000
2001 // Do the 'common' stuff
2002 add_region_to_incremental_cset_common(hr);
2003
2004 // Add the region at the left hand side
2005 hr->set_next_in_collection_set(_inc_cset_head);
2006 if (_inc_cset_head == NULL) {
2007 assert(_inc_cset_tail == NULL, "Invariant");
2008 _inc_cset_tail = hr;
2009 }
2010 _inc_cset_head = hr;
2011 }
2012
2013 #ifndef PRODUCT
2014 void G1CollectorPolicy::print_collection_set(HeapRegion* list_head, outputStream* st) {
2015 assert(list_head == inc_cset_head() || list_head == collection_set(), "must be");
2016
2017 st->print_cr("\nCollection_set:");
2018 HeapRegion* csr = list_head;
2019 while (csr != NULL) {
2020 HeapRegion* next = csr->next_in_collection_set();
2021 assert(csr->in_collection_set(), "bad CS");
2022 st->print_cr(" " HR_FORMAT ", P: " PTR_FORMAT "N: " PTR_FORMAT ", age: %4d",
2023 HR_FORMAT_PARAMS(csr),
2024 p2i(csr->prev_top_at_mark_start()), p2i(csr->next_top_at_mark_start()),
2025 csr->age_in_surv_rate_group_cond());
2026 csr = next;
2027 }
2028 }
2029 #endif // !PRODUCT
2030
2031 double G1CollectorPolicy::reclaimable_bytes_perc(size_t reclaimable_bytes) const {
2032 // Returns the given amount of reclaimable bytes (that represents
2033 // the amount of reclaimable space still to be collected) as a
2034 // percentage of the current heap capacity.
2035 size_t capacity_bytes = _g1->capacity();
2036 return (double) reclaimable_bytes * 100.0 / (double) capacity_bytes;
2037 }
2038
2039 void G1CollectorPolicy::maybe_start_marking() {
2040 if (need_to_start_conc_mark("end of GC")) {
2041 // Note: this might have already been set, if during the last
2042 // pause we decided to start a cycle but at the beginning of
2043 // this pause we decided to postpone it. That's OK.
2044 collector_state()->set_initiate_conc_mark_if_possible(true);
2045 }
2046 }
2047
2048 G1CollectorPolicy::PauseKind G1CollectorPolicy::young_gc_pause_kind() const {
2049 assert(!collector_state()->full_collection(), "must be");
2050 if (collector_state()->during_initial_mark_pause()) {
2051 assert(collector_state()->last_gc_was_young(), "must be");
2052 assert(!collector_state()->last_young_gc(), "must be");
2053 return InitialMarkGC;
2054 } else if (collector_state()->last_young_gc()) {
2055 assert(!collector_state()->during_initial_mark_pause(), "must be");
2056 assert(collector_state()->last_gc_was_young(), "must be");
2057 return LastYoungGC;
2058 } else if (!collector_state()->last_gc_was_young()) {
2059 assert(!collector_state()->during_initial_mark_pause(), "must be");
2060 assert(!collector_state()->last_young_gc(), "must be");
2061 return MixedGC;
2062 } else {
2063 assert(collector_state()->last_gc_was_young(), "must be");
2064 assert(!collector_state()->during_initial_mark_pause(), "must be");
2065 assert(!collector_state()->last_young_gc(), "must be");
2066 return YoungOnlyGC;
2067 }
2068 }
2069
2070 void G1CollectorPolicy::record_pause(PauseKind kind, double start, double end) {
2071 // Manage the MMU tracker. For some reason it ignores Full GCs.
2072 if (kind != FullGC) {
2073 _mmu_tracker->add_pause(start, end);
2074 }
2075 // Manage the mutator time tracking from initial mark to first mixed gc.
2076 switch (kind) {
2077 case FullGC:
2078 abort_time_to_mixed_tracking();
2079 break;
2080 case Cleanup:
2081 case Remark:
2082 case YoungOnlyGC:
2083 case LastYoungGC:
2084 _initial_mark_to_mixed.add_pause(end - start);
2085 break;
2086 case InitialMarkGC:
2087 _initial_mark_to_mixed.record_initial_mark_end(end);
2088 break;
2089 case MixedGC:
2090 _initial_mark_to_mixed.record_mixed_gc_start(start);
2091 break;
2092 default:
2093 ShouldNotReachHere();
2094 }
2095 }
2096
2097 void G1CollectorPolicy::abort_time_to_mixed_tracking() {
2098 _initial_mark_to_mixed.reset();
2099 }
2100
2101 bool G1CollectorPolicy::next_gc_should_be_mixed(const char* true_action_str,
2102 const char* false_action_str) const {
2103 if (cset_chooser()->is_empty()) {
2104 log_debug(gc, ergo)("%s (candidate old regions not available)", false_action_str);
2105 return false;
2106 }
2107
2108 // Is the amount of uncollected reclaimable space above G1HeapWastePercent?
2109 size_t reclaimable_bytes = cset_chooser()->remaining_reclaimable_bytes();
2110 double reclaimable_perc = reclaimable_bytes_perc(reclaimable_bytes);
2111 double threshold = (double) G1HeapWastePercent;
2112 if (reclaimable_perc <= threshold) {
2113 log_debug(gc, ergo)("%s (reclaimable percentage not over threshold). candidate old regions: %u reclaimable: " SIZE_FORMAT " (%1.2f) threshold: " UINTX_FORMAT,
2114 false_action_str, cset_chooser()->remaining_regions(), reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
2115 return false;
2116 }
2117 log_debug(gc, ergo)("%s (candidate old regions available). candidate old regions: %u reclaimable: " SIZE_FORMAT " (%1.2f) threshold: " UINTX_FORMAT,
2118 true_action_str, cset_chooser()->remaining_regions(), reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
2119 return true;
2120 }
2121
2122 uint G1CollectorPolicy::calc_min_old_cset_length() const {
2123 // The min old CSet region bound is based on the maximum desired
2124 // number of mixed GCs after a cycle. I.e., even if some old regions
2125 // look expensive, we should add them to the CSet anyway to make
2126 // sure we go through the available old regions in no more than the
2127 // maximum desired number of mixed GCs.
2128 //
2129 // The calculation is based on the number of marked regions we added
2130 // to the CSet chooser in the first place, not how many remain, so
2131 // that the result is the same during all mixed GCs that follow a cycle.
2132
2133 const size_t region_num = (size_t) cset_chooser()->length();
2134 const size_t gc_num = (size_t) MAX2(G1MixedGCCountTarget, (uintx) 1);
2135 size_t result = region_num / gc_num;
2136 // emulate ceiling
2137 if (result * gc_num < region_num) {
2138 result += 1;
2139 }
2140 return (uint) result;
2141 }
2142
2143 uint G1CollectorPolicy::calc_max_old_cset_length() const {
2144 // The max old CSet region bound is based on the threshold expressed
2145 // as a percentage of the heap size. I.e., it should bound the
2146 // number of old regions added to the CSet irrespective of how many
2147 // of them are available.
2148
2149 const G1CollectedHeap* g1h = G1CollectedHeap::heap();
2150 const size_t region_num = g1h->num_regions();
2151 const size_t perc = (size_t) G1OldCSetRegionThresholdPercent;
2152 size_t result = region_num * perc / 100;
2153 // emulate ceiling
2154 if (100 * result < region_num * perc) {
2155 result += 1;
2156 }
2157 return (uint) result;
2158 }
2159
2160
2161 double G1CollectorPolicy::finalize_young_cset_part(double target_pause_time_ms) {
2162 double young_start_time_sec = os::elapsedTime();
2163
2164 YoungList* young_list = _g1->young_list();
2165 finalize_incremental_cset_building();
2166
2167 guarantee(target_pause_time_ms > 0.0,
2168 "target_pause_time_ms = %1.6lf should be positive", target_pause_time_ms);
2169 guarantee(_collection_set == NULL, "Precondition");
2170
2171 double base_time_ms = predict_base_elapsed_time_ms(_pending_cards);
2172 double time_remaining_ms = MAX2(target_pause_time_ms - base_time_ms, 0.0);
2173
2174 log_trace(gc, ergo, cset)("Start choosing CSet. pending cards: " SIZE_FORMAT " predicted base time: %1.2fms remaining time: %1.2fms target pause time: %1.2fms",
2175 _pending_cards, base_time_ms, time_remaining_ms, target_pause_time_ms);
2176
2177 collector_state()->set_last_gc_was_young(collector_state()->gcs_are_young());
2178
2179 if (collector_state()->last_gc_was_young()) {
2180 _trace_young_gen_time_data.increment_young_collection_count();
2181 } else {
2182 _trace_young_gen_time_data.increment_mixed_collection_count();
2183 }
2184
2185 // The young list is laid with the survivor regions from the previous
2186 // pause are appended to the RHS of the young list, i.e.
2187 // [Newly Young Regions ++ Survivors from last pause].
2188
2189 uint survivor_region_length = young_list->survivor_length();
2190 uint eden_region_length = young_list->eden_length();
2191 init_cset_region_lengths(eden_region_length, survivor_region_length);
2192
2193 HeapRegion* hr = young_list->first_survivor_region();
2194 while (hr != NULL) {
2195 assert(hr->is_survivor(), "badly formed young list");
2196 // There is a convention that all the young regions in the CSet
2197 // are tagged as "eden", so we do this for the survivors here. We
2198 // use the special set_eden_pre_gc() as it doesn't check that the
2199 // region is free (which is not the case here).
2200 hr->set_eden_pre_gc();
2201 hr = hr->get_next_young_region();
2202 }
2203
2204 // Clear the fields that point to the survivor list - they are all young now.
2205 young_list->clear_survivors();
2206
2207 _collection_set = _inc_cset_head;
2208 _collection_set_bytes_used_before = _inc_cset_bytes_used_before;
2209 time_remaining_ms = MAX2(time_remaining_ms - _inc_cset_predicted_elapsed_time_ms, 0.0);
2210
2211 log_trace(gc, ergo, cset)("Add young regions to CSet. eden: %u regions, survivors: %u regions, predicted young region time: %1.2fms, target pause time: %1.2fms",
2212 eden_region_length, survivor_region_length, _inc_cset_predicted_elapsed_time_ms, target_pause_time_ms);
2213
2214 // The number of recorded young regions is the incremental
2215 // collection set's current size
2216 set_recorded_rs_lengths(_inc_cset_recorded_rs_lengths);
2217
2218 double young_end_time_sec = os::elapsedTime();
2219 phase_times()->record_young_cset_choice_time_ms((young_end_time_sec - young_start_time_sec) * 1000.0);
2220
2221 return time_remaining_ms;
2222 }
2223
2224 void G1CollectorPolicy::finalize_old_cset_part(double time_remaining_ms) {
2225 double non_young_start_time_sec = os::elapsedTime();
2226 double predicted_old_time_ms = 0.0;
2227
2228
2229 if (!collector_state()->gcs_are_young()) {
2230 cset_chooser()->verify();
2231 const uint min_old_cset_length = calc_min_old_cset_length();
2232 const uint max_old_cset_length = calc_max_old_cset_length();
2233
2234 uint expensive_region_num = 0;
2235 bool check_time_remaining = adaptive_young_list_length();
2236
2237 HeapRegion* hr = cset_chooser()->peek();
2238 while (hr != NULL) {
2239 if (old_cset_region_length() >= max_old_cset_length) {
2240 // Added maximum number of old regions to the CSet.
2241 log_debug(gc, ergo, cset)("Finish adding old regions to CSet (old CSet region num reached max). old %u regions, max %u regions",
2242 old_cset_region_length(), max_old_cset_length);
2243 break;
2244 }
2245
2246
2247 // Stop adding regions if the remaining reclaimable space is
2248 // not above G1HeapWastePercent.
2249 size_t reclaimable_bytes = cset_chooser()->remaining_reclaimable_bytes();
2250 double reclaimable_perc = reclaimable_bytes_perc(reclaimable_bytes);
2251 double threshold = (double) G1HeapWastePercent;
2252 if (reclaimable_perc <= threshold) {
2253 // We've added enough old regions that the amount of uncollected
2254 // reclaimable space is at or below the waste threshold. Stop
2255 // adding old regions to the CSet.
2256 log_debug(gc, ergo, cset)("Finish adding old regions to CSet (reclaimable percentage not over threshold). "
2257 "old %u regions, max %u regions, reclaimable: " SIZE_FORMAT "B (%1.2f%%) threshold: " UINTX_FORMAT "%%",
2258 old_cset_region_length(), max_old_cset_length, reclaimable_bytes, reclaimable_perc, G1HeapWastePercent);
2259 break;
2260 }
2261
2262 double predicted_time_ms = predict_region_elapsed_time_ms(hr, collector_state()->gcs_are_young());
2263 if (check_time_remaining) {
2264 if (predicted_time_ms > time_remaining_ms) {
2265 // Too expensive for the current CSet.
2266
2267 if (old_cset_region_length() >= min_old_cset_length) {
2268 // We have added the minimum number of old regions to the CSet,
2269 // we are done with this CSet.
2270 log_debug(gc, ergo, cset)("Finish adding old regions to CSet (predicted time is too high). "
2271 "predicted time: %1.2fms, remaining time: %1.2fms old %u regions, min %u regions",
2272 predicted_time_ms, time_remaining_ms, old_cset_region_length(), min_old_cset_length);
2273 break;
2274 }
2275
2276 // We'll add it anyway given that we haven't reached the
2277 // minimum number of old regions.
2278 expensive_region_num += 1;
2279 }
2280 } else {
2281 if (old_cset_region_length() >= min_old_cset_length) {
2282 // In the non-auto-tuning case, we'll finish adding regions
2283 // to the CSet if we reach the minimum.
2284
2285 log_debug(gc, ergo, cset)("Finish adding old regions to CSet (old CSet region num reached min). old %u regions, min %u regions",
2286 old_cset_region_length(), min_old_cset_length);
2287 break;
2288 }
2289 }
2290
2291 // We will add this region to the CSet.
2292 time_remaining_ms = MAX2(time_remaining_ms - predicted_time_ms, 0.0);
2293 predicted_old_time_ms += predicted_time_ms;
2294 cset_chooser()->pop(); // already have region via peek()
2295 _g1->old_set_remove(hr);
2296 add_old_region_to_cset(hr);
2297
2298 hr = cset_chooser()->peek();
2299 }
2300 if (hr == NULL) {
2301 log_debug(gc, ergo, cset)("Finish adding old regions to CSet (candidate old regions not available)");
2302 }
2303
2304 if (expensive_region_num > 0) {
2305 // We print the information once here at the end, predicated on
2306 // whether we added any apparently expensive regions or not, to
2307 // avoid generating output per region.
2308 log_debug(gc, ergo, cset)("Added expensive regions to CSet (old CSet region num not reached min)."
2309 "old %u regions, expensive: %u regions, min %u regions, remaining time: %1.2fms",
2310 old_cset_region_length(), expensive_region_num, min_old_cset_length, time_remaining_ms);
2311 }
2312
2313 cset_chooser()->verify();
2314 }
2315
2316 stop_incremental_cset_building();
2317
2318 log_debug(gc, ergo, cset)("Finish choosing CSet. old %u regions, predicted old region time: %1.2fms, time remaining: %1.2f",
2319 old_cset_region_length(), predicted_old_time_ms, time_remaining_ms);
2320
2321 double non_young_end_time_sec = os::elapsedTime();
2322 phase_times()->record_non_young_cset_choice_time_ms((non_young_end_time_sec - non_young_start_time_sec) * 1000.0);
2323 }
2324
2325 void TraceYoungGenTimeData::record_start_collection(double time_to_stop_the_world_ms) {
2326 if(TraceYoungGenTime) {
2327 _all_stop_world_times_ms.add(time_to_stop_the_world_ms);
2328 }
2329 }
2330
2331 void TraceYoungGenTimeData::record_yield_time(double yield_time_ms) {
2332 if(TraceYoungGenTime) {
2333 _all_yield_times_ms.add(yield_time_ms);
2334 }
2335 }
2336
2337 void TraceYoungGenTimeData::record_end_collection(double pause_time_ms, G1GCPhaseTimes* phase_times) {
2338 if(TraceYoungGenTime) {
2339 _total.add(pause_time_ms);
2340 _other.add(pause_time_ms - phase_times->accounted_time_ms());
2341 _root_region_scan_wait.add(phase_times->root_region_scan_wait_time_ms());
2342 _parallel.add(phase_times->cur_collection_par_time_ms());
2343 _ext_root_scan.add(phase_times->average_time_ms(G1GCPhaseTimes::ExtRootScan));
2344 _satb_filtering.add(phase_times->average_time_ms(G1GCPhaseTimes::SATBFiltering));
2345 _update_rs.add(phase_times->average_time_ms(G1GCPhaseTimes::UpdateRS));
2346 _scan_rs.add(phase_times->average_time_ms(G1GCPhaseTimes::ScanRS));
2347 _obj_copy.add(phase_times->average_time_ms(G1GCPhaseTimes::ObjCopy));
2348 _termination.add(phase_times->average_time_ms(G1GCPhaseTimes::Termination));
2349
2350 double parallel_known_time = phase_times->average_time_ms(G1GCPhaseTimes::ExtRootScan) +
2351 phase_times->average_time_ms(G1GCPhaseTimes::SATBFiltering) +
2352 phase_times->average_time_ms(G1GCPhaseTimes::UpdateRS) +
2353 phase_times->average_time_ms(G1GCPhaseTimes::ScanRS) +
2354 phase_times->average_time_ms(G1GCPhaseTimes::ObjCopy) +
2355 phase_times->average_time_ms(G1GCPhaseTimes::Termination);
2356
2357 double parallel_other_time = phase_times->cur_collection_par_time_ms() - parallel_known_time;
2358 _parallel_other.add(parallel_other_time);
2359 _clear_ct.add(phase_times->cur_clear_ct_time_ms());
2360 }
2361 }
2362
2363 void TraceYoungGenTimeData::increment_young_collection_count() {
2364 if(TraceYoungGenTime) {
2365 ++_young_pause_num;
2366 }
2367 }
2368
2369 void TraceYoungGenTimeData::increment_mixed_collection_count() {
2370 if(TraceYoungGenTime) {
2371 ++_mixed_pause_num;
2372 }
2373 }
2374
2375 void TraceYoungGenTimeData::print_summary(const char* str,
2376 const NumberSeq* seq) const {
2377 double sum = seq->sum();
2378 tty->print_cr("%-27s = %8.2lf s (avg = %8.2lf ms)",
2379 str, sum / 1000.0, seq->avg());
2380 }
2381
2382 void TraceYoungGenTimeData::print_summary_sd(const char* str,
2383 const NumberSeq* seq) const {
2384 print_summary(str, seq);
2385 tty->print_cr("%45s = %5d, std dev = %8.2lf ms, max = %8.2lf ms)",
2386 "(num", seq->num(), seq->sd(), seq->maximum());
2387 }
2388
2389 void TraceYoungGenTimeData::print() const {
2390 if (!TraceYoungGenTime) {
2391 return;
2392 }
2393
2394 tty->print_cr("ALL PAUSES");
2395 print_summary_sd(" Total", &_total);
2396 tty->cr();
2397 tty->cr();
2398 tty->print_cr(" Young GC Pauses: %8d", _young_pause_num);
2399 tty->print_cr(" Mixed GC Pauses: %8d", _mixed_pause_num);
2400 tty->cr();
2401
2402 tty->print_cr("EVACUATION PAUSES");
2403
2404 if (_young_pause_num == 0 && _mixed_pause_num == 0) {
2405 tty->print_cr("none");
2406 } else {
2407 print_summary_sd(" Evacuation Pauses", &_total);
2408 print_summary(" Root Region Scan Wait", &_root_region_scan_wait);
2409 print_summary(" Parallel Time", &_parallel);
2410 print_summary(" Ext Root Scanning", &_ext_root_scan);
2411 print_summary(" SATB Filtering", &_satb_filtering);
2412 print_summary(" Update RS", &_update_rs);
2413 print_summary(" Scan RS", &_scan_rs);
2414 print_summary(" Object Copy", &_obj_copy);
2415 print_summary(" Termination", &_termination);
2416 print_summary(" Parallel Other", &_parallel_other);
2417 print_summary(" Clear CT", &_clear_ct);
2418 print_summary(" Other", &_other);
2419 }
2420 tty->cr();
2421
2422 tty->print_cr("MISC");
2423 print_summary_sd(" Stop World", &_all_stop_world_times_ms);
2424 print_summary_sd(" Yields", &_all_yield_times_ms);
2425 }
2426
2427 void TraceOldGenTimeData::record_full_collection(double full_gc_time_ms) {
2428 if (TraceOldGenTime) {
2429 _all_full_gc_times.add(full_gc_time_ms);
2430 }
2431 }
2432
2433 void TraceOldGenTimeData::print() const {
2434 if (!TraceOldGenTime) {
2435 return;
2436 }
2437
2438 if (_all_full_gc_times.num() > 0) {
2439 tty->print("\n%4d full_gcs: total time = %8.2f s",
2440 _all_full_gc_times.num(),
2441 _all_full_gc_times.sum() / 1000.0);
2442 tty->print_cr(" (avg = %8.2fms).", _all_full_gc_times.avg());
2443 tty->print_cr(" [std. dev = %8.2f ms, max = %8.2f ms]",
2444 _all_full_gc_times.sd(),
2445 _all_full_gc_times.maximum());
2446 }
2447 }