1 /*
2 * Copyright (c) 2001, 2012, 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_implementation/concurrentMarkSweep/cmsLockVerifier.hpp"
27 #include "gc_implementation/concurrentMarkSweep/compactibleFreeListSpace.hpp"
28 #include "gc_implementation/concurrentMarkSweep/concurrentMarkSweepGeneration.inline.hpp"
29 #include "gc_implementation/concurrentMarkSweep/concurrentMarkSweepThread.hpp"
30 #include "gc_implementation/shared/liveRange.hpp"
31 #include "gc_implementation/shared/spaceDecorator.hpp"
32 #include "gc_interface/collectedHeap.hpp"
33 #include "memory/allocation.inline.hpp"
34 #include "memory/blockOffsetTable.inline.hpp"
35 #include "memory/resourceArea.hpp"
36 #include "memory/universe.inline.hpp"
37 #include "oops/oop.inline.hpp"
38 #include "runtime/globals.hpp"
39 #include "runtime/handles.inline.hpp"
40 #include "runtime/init.hpp"
41 #include "runtime/java.hpp"
42 #include "runtime/vmThread.hpp"
43 #include "utilities/copy.hpp"
44
45 /////////////////////////////////////////////////////////////////////////
46 //// CompactibleFreeListSpace
47 /////////////////////////////////////////////////////////////////////////
48
49 // highest ranked free list lock rank
50 int CompactibleFreeListSpace::_lockRank = Mutex::leaf + 3;
51
52 // Defaults are 0 so things will break badly if incorrectly initialized.
53 size_t CompactibleFreeListSpace::IndexSetStart = 0;
54 size_t CompactibleFreeListSpace::IndexSetStride = 0;
55
56 size_t MinChunkSize = 0;
57
58 void CompactibleFreeListSpace::set_cms_values() {
59 // Set CMS global values
60 assert(MinChunkSize == 0, "already set");
61
62 // MinChunkSize should be a multiple of MinObjAlignment and be large enough
63 // for chunks to contain a FreeChunk.
64 size_t min_chunk_size_in_bytes = align_size_up(sizeof(FreeChunk), MinObjAlignmentInBytes);
65 MinChunkSize = min_chunk_size_in_bytes / BytesPerWord;
66
67 assert(IndexSetStart == 0 && IndexSetStride == 0, "already set");
68 IndexSetStart = MinChunkSize;
69 IndexSetStride = MinObjAlignment;
70 }
71
72 // Constructor
73 CompactibleFreeListSpace::CompactibleFreeListSpace(BlockOffsetSharedArray* bs,
74 MemRegion mr, bool use_adaptive_freelists,
75 FreeBlockDictionary<FreeChunk>::DictionaryChoice dictionaryChoice) :
76 _dictionaryChoice(dictionaryChoice),
77 _adaptive_freelists(use_adaptive_freelists),
78 _bt(bs, mr),
79 // free list locks are in the range of values taken by _lockRank
80 // This range currently is [_leaf+2, _leaf+3]
81 // Note: this requires that CFLspace c'tors
82 // are called serially in the order in which the locks are
83 // are acquired in the program text. This is true today.
84 _freelistLock(_lockRank--, "CompactibleFreeListSpace._lock", true),
85 _parDictionaryAllocLock(Mutex::leaf - 1, // == rank(ExpandHeap_lock) - 1
86 "CompactibleFreeListSpace._dict_par_lock", true),
87 _rescan_task_size(CardTableModRefBS::card_size_in_words * BitsPerWord *
88 CMSRescanMultiple),
89 _marking_task_size(CardTableModRefBS::card_size_in_words * BitsPerWord *
90 CMSConcMarkMultiple),
91 _collector(NULL)
92 {
93 assert(sizeof(FreeChunk) / BytesPerWord <= MinChunkSize,
94 "FreeChunk is larger than expected");
95 _bt.set_space(this);
96 initialize(mr, SpaceDecorator::Clear, SpaceDecorator::Mangle);
97 // We have all of "mr", all of which we place in the dictionary
98 // as one big chunk. We'll need to decide here which of several
99 // possible alternative dictionary implementations to use. For
100 // now the choice is easy, since we have only one working
101 // implementation, namely, the simple binary tree (splaying
102 // temporarily disabled).
103 switch (dictionaryChoice) {
104 case FreeBlockDictionary<FreeChunk>::dictionarySplayTree:
105 case FreeBlockDictionary<FreeChunk>::dictionarySkipList:
106 default:
107 warning("dictionaryChoice: selected option not understood; using"
108 " default BinaryTreeDictionary implementation instead.");
109 case FreeBlockDictionary<FreeChunk>::dictionaryBinaryTree:
110 _dictionary = new BinaryTreeDictionary<FreeChunk>(mr, use_adaptive_freelists);
111 break;
112 }
113 assert(_dictionary != NULL, "CMS dictionary initialization");
114 // The indexed free lists are initially all empty and are lazily
115 // filled in on demand. Initialize the array elements to NULL.
116 initializeIndexedFreeListArray();
117
118 // Not using adaptive free lists assumes that allocation is first
119 // from the linAB's. Also a cms perm gen which can be compacted
120 // has to have the klass's klassKlass allocated at a lower
121 // address in the heap than the klass so that the klassKlass is
122 // moved to its new location before the klass is moved.
123 // Set the _refillSize for the linear allocation blocks
124 if (!use_adaptive_freelists) {
125 FreeChunk* fc = _dictionary->get_chunk(mr.word_size());
126 // The small linAB initially has all the space and will allocate
127 // a chunk of any size.
128 HeapWord* addr = (HeapWord*) fc;
129 _smallLinearAllocBlock.set(addr, fc->size() ,
130 1024*SmallForLinearAlloc, fc->size());
131 // Note that _unallocated_block is not updated here.
132 // Allocations from the linear allocation block should
133 // update it.
134 } else {
135 _smallLinearAllocBlock.set(0, 0, 1024*SmallForLinearAlloc,
136 SmallForLinearAlloc);
137 }
138 // CMSIndexedFreeListReplenish should be at least 1
139 CMSIndexedFreeListReplenish = MAX2((uintx)1, CMSIndexedFreeListReplenish);
140 _promoInfo.setSpace(this);
141 if (UseCMSBestFit) {
142 _fitStrategy = FreeBlockBestFitFirst;
143 } else {
144 _fitStrategy = FreeBlockStrategyNone;
145 }
146 check_free_list_consistency();
147
148 // Initialize locks for parallel case.
149
150 if (CollectedHeap::use_parallel_gc_threads()) {
151 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
152 _indexedFreeListParLocks[i] = new Mutex(Mutex::leaf - 1, // == ExpandHeap_lock - 1
153 "a freelist par lock",
154 true);
155 if (_indexedFreeListParLocks[i] == NULL)
156 vm_exit_during_initialization("Could not allocate a par lock");
157 DEBUG_ONLY(
158 _indexedFreeList[i].set_protecting_lock(_indexedFreeListParLocks[i]);
159 )
160 }
161 _dictionary->set_par_lock(&_parDictionaryAllocLock);
162 }
163 }
164
165 // Like CompactibleSpace forward() but always calls cross_threshold() to
166 // update the block offset table. Removed initialize_threshold call because
167 // CFLS does not use a block offset array for contiguous spaces.
168 HeapWord* CompactibleFreeListSpace::forward(oop q, size_t size,
169 CompactPoint* cp, HeapWord* compact_top) {
170 // q is alive
171 // First check if we should switch compaction space
172 assert(this == cp->space, "'this' should be current compaction space.");
173 size_t compaction_max_size = pointer_delta(end(), compact_top);
174 assert(adjustObjectSize(size) == cp->space->adjust_object_size_v(size),
175 "virtual adjustObjectSize_v() method is not correct");
176 size_t adjusted_size = adjustObjectSize(size);
177 assert(compaction_max_size >= MinChunkSize || compaction_max_size == 0,
178 "no small fragments allowed");
179 assert(minimum_free_block_size() == MinChunkSize,
180 "for de-virtualized reference below");
181 // Can't leave a nonzero size, residual fragment smaller than MinChunkSize
182 if (adjusted_size + MinChunkSize > compaction_max_size &&
183 adjusted_size != compaction_max_size) {
184 do {
185 // switch to next compaction space
186 cp->space->set_compaction_top(compact_top);
187 cp->space = cp->space->next_compaction_space();
188 if (cp->space == NULL) {
189 cp->gen = GenCollectedHeap::heap()->prev_gen(cp->gen);
190 assert(cp->gen != NULL, "compaction must succeed");
191 cp->space = cp->gen->first_compaction_space();
192 assert(cp->space != NULL, "generation must have a first compaction space");
193 }
194 compact_top = cp->space->bottom();
195 cp->space->set_compaction_top(compact_top);
196 // The correct adjusted_size may not be the same as that for this method
197 // (i.e., cp->space may no longer be "this" so adjust the size again.
198 // Use the virtual method which is not used above to save the virtual
199 // dispatch.
200 adjusted_size = cp->space->adjust_object_size_v(size);
201 compaction_max_size = pointer_delta(cp->space->end(), compact_top);
202 assert(cp->space->minimum_free_block_size() == 0, "just checking");
203 } while (adjusted_size > compaction_max_size);
204 }
205
206 // store the forwarding pointer into the mark word
207 if ((HeapWord*)q != compact_top) {
208 q->forward_to(oop(compact_top));
209 assert(q->is_gc_marked(), "encoding the pointer should preserve the mark");
210 } else {
211 // if the object isn't moving we can just set the mark to the default
212 // mark and handle it specially later on.
213 q->init_mark();
214 assert(q->forwardee() == NULL, "should be forwarded to NULL");
215 }
216
217 VALIDATE_MARK_SWEEP_ONLY(MarkSweep::register_live_oop(q, adjusted_size));
218 compact_top += adjusted_size;
219
220 // we need to update the offset table so that the beginnings of objects can be
221 // found during scavenge. Note that we are updating the offset table based on
222 // where the object will be once the compaction phase finishes.
223
224 // Always call cross_threshold(). A contiguous space can only call it when
225 // the compaction_top exceeds the current threshold but not for an
226 // non-contiguous space.
227 cp->threshold =
228 cp->space->cross_threshold(compact_top - adjusted_size, compact_top);
229 return compact_top;
230 }
231
232 // A modified copy of OffsetTableContigSpace::cross_threshold() with _offsets -> _bt
233 // and use of single_block instead of alloc_block. The name here is not really
234 // appropriate - maybe a more general name could be invented for both the
235 // contiguous and noncontiguous spaces.
236
237 HeapWord* CompactibleFreeListSpace::cross_threshold(HeapWord* start, HeapWord* the_end) {
238 _bt.single_block(start, the_end);
239 return end();
240 }
241
242 // Initialize them to NULL.
243 void CompactibleFreeListSpace::initializeIndexedFreeListArray() {
244 for (size_t i = 0; i < IndexSetSize; i++) {
245 // Note that on platforms where objects are double word aligned,
246 // the odd array elements are not used. It is convenient, however,
247 // to map directly from the object size to the array element.
248 _indexedFreeList[i].reset(IndexSetSize);
249 _indexedFreeList[i].set_size(i);
250 assert(_indexedFreeList[i].count() == 0, "reset check failed");
251 assert(_indexedFreeList[i].head() == NULL, "reset check failed");
252 assert(_indexedFreeList[i].tail() == NULL, "reset check failed");
253 assert(_indexedFreeList[i].hint() == IndexSetSize, "reset check failed");
254 }
255 }
256
257 void CompactibleFreeListSpace::resetIndexedFreeListArray() {
258 for (size_t i = 1; i < IndexSetSize; i++) {
259 assert(_indexedFreeList[i].size() == (size_t) i,
260 "Indexed free list sizes are incorrect");
261 _indexedFreeList[i].reset(IndexSetSize);
262 assert(_indexedFreeList[i].count() == 0, "reset check failed");
263 assert(_indexedFreeList[i].head() == NULL, "reset check failed");
264 assert(_indexedFreeList[i].tail() == NULL, "reset check failed");
265 assert(_indexedFreeList[i].hint() == IndexSetSize, "reset check failed");
266 }
267 }
268
269 void CompactibleFreeListSpace::reset(MemRegion mr) {
270 resetIndexedFreeListArray();
271 dictionary()->reset();
272 if (BlockOffsetArrayUseUnallocatedBlock) {
273 assert(end() == mr.end(), "We are compacting to the bottom of CMS gen");
274 // Everything's allocated until proven otherwise.
275 _bt.set_unallocated_block(end());
276 }
277 if (!mr.is_empty()) {
278 assert(mr.word_size() >= MinChunkSize, "Chunk size is too small");
279 _bt.single_block(mr.start(), mr.word_size());
280 FreeChunk* fc = (FreeChunk*) mr.start();
281 fc->set_size(mr.word_size());
282 if (mr.word_size() >= IndexSetSize ) {
283 returnChunkToDictionary(fc);
284 } else {
285 _bt.verify_not_unallocated((HeapWord*)fc, fc->size());
286 _indexedFreeList[mr.word_size()].return_chunk_at_head(fc);
287 }
288 }
289 _promoInfo.reset();
290 _smallLinearAllocBlock._ptr = NULL;
291 _smallLinearAllocBlock._word_size = 0;
292 }
293
294 void CompactibleFreeListSpace::reset_after_compaction() {
295 // Reset the space to the new reality - one free chunk.
296 MemRegion mr(compaction_top(), end());
297 reset(mr);
298 // Now refill the linear allocation block(s) if possible.
299 if (_adaptive_freelists) {
300 refillLinearAllocBlocksIfNeeded();
301 } else {
302 // Place as much of mr in the linAB as we can get,
303 // provided it was big enough to go into the dictionary.
304 FreeChunk* fc = dictionary()->find_largest_dict();
305 if (fc != NULL) {
306 assert(fc->size() == mr.word_size(),
307 "Why was the chunk broken up?");
308 removeChunkFromDictionary(fc);
309 HeapWord* addr = (HeapWord*) fc;
310 _smallLinearAllocBlock.set(addr, fc->size() ,
311 1024*SmallForLinearAlloc, fc->size());
312 // Note that _unallocated_block is not updated here.
313 }
314 }
315 }
316
317 // Walks the entire dictionary, returning a coterminal
318 // chunk, if it exists. Use with caution since it involves
319 // a potentially complete walk of a potentially large tree.
320 FreeChunk* CompactibleFreeListSpace::find_chunk_at_end() {
321
322 assert_lock_strong(&_freelistLock);
323
324 return dictionary()->find_chunk_ends_at(end());
325 }
326
327
328 #ifndef PRODUCT
329 void CompactibleFreeListSpace::initializeIndexedFreeListArrayReturnedBytes() {
330 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
331 _indexedFreeList[i].allocation_stats()->set_returned_bytes(0);
332 }
333 }
334
335 size_t CompactibleFreeListSpace::sumIndexedFreeListArrayReturnedBytes() {
336 size_t sum = 0;
337 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
338 sum += _indexedFreeList[i].allocation_stats()->returned_bytes();
339 }
340 return sum;
341 }
342
343 size_t CompactibleFreeListSpace::totalCountInIndexedFreeLists() const {
344 size_t count = 0;
345 for (size_t i = IndexSetStart; i < IndexSetSize; i++) {
346 debug_only(
347 ssize_t total_list_count = 0;
348 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL;
349 fc = fc->next()) {
350 total_list_count++;
351 }
352 assert(total_list_count == _indexedFreeList[i].count(),
353 "Count in list is incorrect");
354 )
355 count += _indexedFreeList[i].count();
356 }
357 return count;
358 }
359
360 size_t CompactibleFreeListSpace::totalCount() {
361 size_t num = totalCountInIndexedFreeLists();
362 num += dictionary()->total_count();
363 if (_smallLinearAllocBlock._word_size != 0) {
364 num++;
365 }
366 return num;
367 }
368 #endif
369
370 bool CompactibleFreeListSpace::is_free_block(const HeapWord* p) const {
371 FreeChunk* fc = (FreeChunk*) p;
372 return fc->is_free();
373 }
374
375 size_t CompactibleFreeListSpace::used() const {
376 return capacity() - free();
377 }
378
379 size_t CompactibleFreeListSpace::free() const {
380 // "MT-safe, but not MT-precise"(TM), if you will: i.e.
381 // if you do this while the structures are in flux you
382 // may get an approximate answer only; for instance
383 // because there is concurrent allocation either
384 // directly by mutators or for promotion during a GC.
385 // It's "MT-safe", however, in the sense that you are guaranteed
386 // not to crash and burn, for instance, because of walking
387 // pointers that could disappear as you were walking them.
388 // The approximation is because the various components
389 // that are read below are not read atomically (and
390 // further the computation of totalSizeInIndexedFreeLists()
391 // is itself a non-atomic computation. The normal use of
392 // this is during a resize operation at the end of GC
393 // and at that time you are guaranteed to get the
394 // correct actual value. However, for instance, this is
395 // also read completely asynchronously by the "perf-sampler"
396 // that supports jvmstat, and you are apt to see the values
397 // flicker in such cases.
398 assert(_dictionary != NULL, "No _dictionary?");
399 return (_dictionary->total_chunk_size(DEBUG_ONLY(freelistLock())) +
400 totalSizeInIndexedFreeLists() +
401 _smallLinearAllocBlock._word_size) * HeapWordSize;
402 }
403
404 size_t CompactibleFreeListSpace::max_alloc_in_words() const {
405 assert(_dictionary != NULL, "No _dictionary?");
406 assert_locked();
407 size_t res = _dictionary->max_chunk_size();
408 res = MAX2(res, MIN2(_smallLinearAllocBlock._word_size,
409 (size_t) SmallForLinearAlloc - 1));
410 // XXX the following could potentially be pretty slow;
411 // should one, pesimally for the rare cases when res
412 // caclulated above is less than IndexSetSize,
413 // just return res calculated above? My reasoning was that
414 // those cases will be so rare that the extra time spent doesn't
415 // really matter....
416 // Note: do not change the loop test i >= res + IndexSetStride
417 // to i > res below, because i is unsigned and res may be zero.
418 for (size_t i = IndexSetSize - 1; i >= res + IndexSetStride;
419 i -= IndexSetStride) {
420 if (_indexedFreeList[i].head() != NULL) {
421 assert(_indexedFreeList[i].count() != 0, "Inconsistent FreeList");
422 return i;
423 }
424 }
425 return res;
426 }
427
428 void LinearAllocBlock::print_on(outputStream* st) const {
429 st->print_cr(" LinearAllocBlock: ptr = " PTR_FORMAT ", word_size = " SIZE_FORMAT
430 ", refillsize = " SIZE_FORMAT ", allocation_size_limit = " SIZE_FORMAT,
431 _ptr, _word_size, _refillSize, _allocation_size_limit);
432 }
433
434 void CompactibleFreeListSpace::print_on(outputStream* st) const {
435 st->print_cr("COMPACTIBLE FREELIST SPACE");
436 st->print_cr(" Space:");
437 Space::print_on(st);
438
439 st->print_cr("promoInfo:");
440 _promoInfo.print_on(st);
441
442 st->print_cr("_smallLinearAllocBlock");
443 _smallLinearAllocBlock.print_on(st);
444
445 // dump_memory_block(_smallLinearAllocBlock->_ptr, 128);
446
447 st->print_cr(" _fitStrategy = %s, _adaptive_freelists = %s",
448 _fitStrategy?"true":"false", _adaptive_freelists?"true":"false");
449 }
450
451 void CompactibleFreeListSpace::print_indexed_free_lists(outputStream* st)
452 const {
453 reportIndexedFreeListStatistics();
454 gclog_or_tty->print_cr("Layout of Indexed Freelists");
455 gclog_or_tty->print_cr("---------------------------");
456 FreeList<FreeChunk>::print_labels_on(st, "size");
457 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
458 _indexedFreeList[i].print_on(gclog_or_tty);
459 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL;
460 fc = fc->next()) {
461 gclog_or_tty->print_cr("\t[" PTR_FORMAT "," PTR_FORMAT ") %s",
462 fc, (HeapWord*)fc + i,
463 fc->cantCoalesce() ? "\t CC" : "");
464 }
465 }
466 }
467
468 void CompactibleFreeListSpace::print_promo_info_blocks(outputStream* st)
469 const {
470 _promoInfo.print_on(st);
471 }
472
473 void CompactibleFreeListSpace::print_dictionary_free_lists(outputStream* st)
474 const {
475 _dictionary->report_statistics();
476 st->print_cr("Layout of Freelists in Tree");
477 st->print_cr("---------------------------");
478 _dictionary->print_free_lists(st);
479 }
480
481 class BlkPrintingClosure: public BlkClosure {
482 const CMSCollector* _collector;
483 const CompactibleFreeListSpace* _sp;
484 const CMSBitMap* _live_bit_map;
485 const bool _post_remark;
486 outputStream* _st;
487 public:
488 BlkPrintingClosure(const CMSCollector* collector,
489 const CompactibleFreeListSpace* sp,
490 const CMSBitMap* live_bit_map,
491 outputStream* st):
492 _collector(collector),
493 _sp(sp),
494 _live_bit_map(live_bit_map),
495 _post_remark(collector->abstract_state() > CMSCollector::FinalMarking),
496 _st(st) { }
497 size_t do_blk(HeapWord* addr);
498 };
499
500 size_t BlkPrintingClosure::do_blk(HeapWord* addr) {
501 size_t sz = _sp->block_size_no_stall(addr, _collector);
502 assert(sz != 0, "Should always be able to compute a size");
503 if (_sp->block_is_obj(addr)) {
504 const bool dead = _post_remark && !_live_bit_map->isMarked(addr);
505 _st->print_cr(PTR_FORMAT ": %s object of size " SIZE_FORMAT "%s",
506 addr,
507 dead ? "dead" : "live",
508 sz,
509 (!dead && CMSPrintObjectsInDump) ? ":" : ".");
510 if (CMSPrintObjectsInDump && !dead) {
511 oop(addr)->print_on(_st);
512 _st->print_cr("--------------------------------------");
513 }
514 } else { // free block
515 _st->print_cr(PTR_FORMAT ": free block of size " SIZE_FORMAT "%s",
516 addr, sz, CMSPrintChunksInDump ? ":" : ".");
517 if (CMSPrintChunksInDump) {
518 ((FreeChunk*)addr)->print_on(_st);
519 _st->print_cr("--------------------------------------");
520 }
521 }
522 return sz;
523 }
524
525 void CompactibleFreeListSpace::dump_at_safepoint_with_locks(CMSCollector* c,
526 outputStream* st) {
527 st->print_cr("\n=========================");
528 st->print_cr("Block layout in CMS Heap:");
529 st->print_cr("=========================");
530 BlkPrintingClosure bpcl(c, this, c->markBitMap(), st);
531 blk_iterate(&bpcl);
532
533 st->print_cr("\n=======================================");
534 st->print_cr("Order & Layout of Promotion Info Blocks");
535 st->print_cr("=======================================");
536 print_promo_info_blocks(st);
537
538 st->print_cr("\n===========================");
539 st->print_cr("Order of Indexed Free Lists");
540 st->print_cr("=========================");
541 print_indexed_free_lists(st);
542
543 st->print_cr("\n=================================");
544 st->print_cr("Order of Free Lists in Dictionary");
545 st->print_cr("=================================");
546 print_dictionary_free_lists(st);
547 }
548
549
550 void CompactibleFreeListSpace::reportFreeListStatistics() const {
551 assert_lock_strong(&_freelistLock);
552 assert(PrintFLSStatistics != 0, "Reporting error");
553 _dictionary->report_statistics();
554 if (PrintFLSStatistics > 1) {
555 reportIndexedFreeListStatistics();
556 size_t total_size = totalSizeInIndexedFreeLists() +
557 _dictionary->total_chunk_size(DEBUG_ONLY(freelistLock()));
558 gclog_or_tty->print(" free=%ld frag=%1.4f\n", total_size, flsFrag());
559 }
560 }
561
562 void CompactibleFreeListSpace::reportIndexedFreeListStatistics() const {
563 assert_lock_strong(&_freelistLock);
564 gclog_or_tty->print("Statistics for IndexedFreeLists:\n"
565 "--------------------------------\n");
566 size_t total_size = totalSizeInIndexedFreeLists();
567 size_t free_blocks = numFreeBlocksInIndexedFreeLists();
568 gclog_or_tty->print("Total Free Space: %d\n", total_size);
569 gclog_or_tty->print("Max Chunk Size: %d\n", maxChunkSizeInIndexedFreeLists());
570 gclog_or_tty->print("Number of Blocks: %d\n", free_blocks);
571 if (free_blocks != 0) {
572 gclog_or_tty->print("Av. Block Size: %d\n", total_size/free_blocks);
573 }
574 }
575
576 size_t CompactibleFreeListSpace::numFreeBlocksInIndexedFreeLists() const {
577 size_t res = 0;
578 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
579 debug_only(
580 ssize_t recount = 0;
581 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL;
582 fc = fc->next()) {
583 recount += 1;
584 }
585 assert(recount == _indexedFreeList[i].count(),
586 "Incorrect count in list");
587 )
588 res += _indexedFreeList[i].count();
589 }
590 return res;
591 }
592
593 size_t CompactibleFreeListSpace::maxChunkSizeInIndexedFreeLists() const {
594 for (size_t i = IndexSetSize - 1; i != 0; i -= IndexSetStride) {
595 if (_indexedFreeList[i].head() != NULL) {
596 assert(_indexedFreeList[i].count() != 0, "Inconsistent FreeList");
597 return (size_t)i;
598 }
599 }
600 return 0;
601 }
602
603 void CompactibleFreeListSpace::set_end(HeapWord* value) {
604 HeapWord* prevEnd = end();
605 assert(prevEnd != value, "unnecessary set_end call");
606 assert(prevEnd == NULL || !BlockOffsetArrayUseUnallocatedBlock || value >= unallocated_block(),
607 "New end is below unallocated block");
608 _end = value;
609 if (prevEnd != NULL) {
610 // Resize the underlying block offset table.
611 _bt.resize(pointer_delta(value, bottom()));
612 if (value <= prevEnd) {
613 assert(!BlockOffsetArrayUseUnallocatedBlock || value >= unallocated_block(),
614 "New end is below unallocated block");
615 } else {
616 // Now, take this new chunk and add it to the free blocks.
617 // Note that the BOT has not yet been updated for this block.
618 size_t newFcSize = pointer_delta(value, prevEnd);
619 // XXX This is REALLY UGLY and should be fixed up. XXX
620 if (!_adaptive_freelists && _smallLinearAllocBlock._ptr == NULL) {
621 // Mark the boundary of the new block in BOT
622 _bt.mark_block(prevEnd, value);
623 // put it all in the linAB
624 if (ParallelGCThreads == 0) {
625 _smallLinearAllocBlock._ptr = prevEnd;
626 _smallLinearAllocBlock._word_size = newFcSize;
627 repairLinearAllocBlock(&_smallLinearAllocBlock);
628 } else { // ParallelGCThreads > 0
629 MutexLockerEx x(parDictionaryAllocLock(),
630 Mutex::_no_safepoint_check_flag);
631 _smallLinearAllocBlock._ptr = prevEnd;
632 _smallLinearAllocBlock._word_size = newFcSize;
633 repairLinearAllocBlock(&_smallLinearAllocBlock);
634 }
635 // Births of chunks put into a LinAB are not recorded. Births
636 // of chunks as they are allocated out of a LinAB are.
637 } else {
638 // Add the block to the free lists, if possible coalescing it
639 // with the last free block, and update the BOT and census data.
640 addChunkToFreeListsAtEndRecordingStats(prevEnd, newFcSize);
641 }
642 }
643 }
644 }
645
646 class FreeListSpace_DCTOC : public Filtering_DCTOC {
647 CompactibleFreeListSpace* _cfls;
648 CMSCollector* _collector;
649 protected:
650 // Override.
651 #define walk_mem_region_with_cl_DECL(ClosureType) \
652 virtual void walk_mem_region_with_cl(MemRegion mr, \
653 HeapWord* bottom, HeapWord* top, \
654 ClosureType* cl); \
655 void walk_mem_region_with_cl_par(MemRegion mr, \
656 HeapWord* bottom, HeapWord* top, \
657 ClosureType* cl); \
658 void walk_mem_region_with_cl_nopar(MemRegion mr, \
659 HeapWord* bottom, HeapWord* top, \
660 ClosureType* cl)
661 walk_mem_region_with_cl_DECL(OopClosure);
662 walk_mem_region_with_cl_DECL(FilteringClosure);
663
664 public:
665 FreeListSpace_DCTOC(CompactibleFreeListSpace* sp,
666 CMSCollector* collector,
667 OopClosure* cl,
668 CardTableModRefBS::PrecisionStyle precision,
669 HeapWord* boundary) :
670 Filtering_DCTOC(sp, cl, precision, boundary),
671 _cfls(sp), _collector(collector) {}
672 };
673
674 // We de-virtualize the block-related calls below, since we know that our
675 // space is a CompactibleFreeListSpace.
676
677 #define FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(ClosureType) \
678 void FreeListSpace_DCTOC::walk_mem_region_with_cl(MemRegion mr, \
679 HeapWord* bottom, \
680 HeapWord* top, \
681 ClosureType* cl) { \
682 bool is_par = SharedHeap::heap()->n_par_threads() > 0; \
683 if (is_par) { \
684 assert(SharedHeap::heap()->n_par_threads() == \
685 SharedHeap::heap()->workers()->active_workers(), "Mismatch"); \
686 walk_mem_region_with_cl_par(mr, bottom, top, cl); \
687 } else { \
688 walk_mem_region_with_cl_nopar(mr, bottom, top, cl); \
689 } \
690 } \
691 void FreeListSpace_DCTOC::walk_mem_region_with_cl_par(MemRegion mr, \
692 HeapWord* bottom, \
693 HeapWord* top, \
694 ClosureType* cl) { \
695 /* Skip parts that are before "mr", in case "block_start" sent us \
696 back too far. */ \
697 HeapWord* mr_start = mr.start(); \
698 size_t bot_size = _cfls->CompactibleFreeListSpace::block_size(bottom); \
699 HeapWord* next = bottom + bot_size; \
700 while (next < mr_start) { \
701 bottom = next; \
702 bot_size = _cfls->CompactibleFreeListSpace::block_size(bottom); \
703 next = bottom + bot_size; \
704 } \
705 \
706 while (bottom < top) { \
707 if (_cfls->CompactibleFreeListSpace::block_is_obj(bottom) && \
708 !_cfls->CompactibleFreeListSpace::obj_allocated_since_save_marks( \
709 oop(bottom)) && \
710 !_collector->CMSCollector::is_dead_obj(oop(bottom))) { \
711 size_t word_sz = oop(bottom)->oop_iterate(cl, mr); \
712 bottom += _cfls->adjustObjectSize(word_sz); \
713 } else { \
714 bottom += _cfls->CompactibleFreeListSpace::block_size(bottom); \
715 } \
716 } \
717 } \
718 void FreeListSpace_DCTOC::walk_mem_region_with_cl_nopar(MemRegion mr, \
719 HeapWord* bottom, \
720 HeapWord* top, \
721 ClosureType* cl) { \
722 /* Skip parts that are before "mr", in case "block_start" sent us \
723 back too far. */ \
724 HeapWord* mr_start = mr.start(); \
725 size_t bot_size = _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \
726 HeapWord* next = bottom + bot_size; \
727 while (next < mr_start) { \
728 bottom = next; \
729 bot_size = _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \
730 next = bottom + bot_size; \
731 } \
732 \
733 while (bottom < top) { \
734 if (_cfls->CompactibleFreeListSpace::block_is_obj_nopar(bottom) && \
735 !_cfls->CompactibleFreeListSpace::obj_allocated_since_save_marks( \
736 oop(bottom)) && \
737 !_collector->CMSCollector::is_dead_obj(oop(bottom))) { \
738 size_t word_sz = oop(bottom)->oop_iterate(cl, mr); \
739 bottom += _cfls->adjustObjectSize(word_sz); \
740 } else { \
741 bottom += _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \
742 } \
743 } \
744 }
745
746 // (There are only two of these, rather than N, because the split is due
747 // only to the introduction of the FilteringClosure, a local part of the
748 // impl of this abstraction.)
749 FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(OopClosure)
750 FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(FilteringClosure)
751
752 DirtyCardToOopClosure*
753 CompactibleFreeListSpace::new_dcto_cl(OopClosure* cl,
754 CardTableModRefBS::PrecisionStyle precision,
755 HeapWord* boundary) {
756 return new FreeListSpace_DCTOC(this, _collector, cl, precision, boundary);
757 }
758
759
760 // Note on locking for the space iteration functions:
761 // since the collector's iteration activities are concurrent with
762 // allocation activities by mutators, absent a suitable mutual exclusion
763 // mechanism the iterators may go awry. For instace a block being iterated
764 // may suddenly be allocated or divided up and part of it allocated and
765 // so on.
766
767 // Apply the given closure to each block in the space.
768 void CompactibleFreeListSpace::blk_iterate_careful(BlkClosureCareful* cl) {
769 assert_lock_strong(freelistLock());
770 HeapWord *cur, *limit;
771 for (cur = bottom(), limit = end(); cur < limit;
772 cur += cl->do_blk_careful(cur));
773 }
774
775 // Apply the given closure to each block in the space.
776 void CompactibleFreeListSpace::blk_iterate(BlkClosure* cl) {
777 assert_lock_strong(freelistLock());
778 HeapWord *cur, *limit;
779 for (cur = bottom(), limit = end(); cur < limit;
780 cur += cl->do_blk(cur));
781 }
782
783 // Apply the given closure to each oop in the space.
784 void CompactibleFreeListSpace::oop_iterate(OopClosure* cl) {
785 assert_lock_strong(freelistLock());
786 HeapWord *cur, *limit;
787 size_t curSize;
788 for (cur = bottom(), limit = end(); cur < limit;
789 cur += curSize) {
790 curSize = block_size(cur);
791 if (block_is_obj(cur)) {
792 oop(cur)->oop_iterate(cl);
793 }
794 }
795 }
796
797 // Apply the given closure to each oop in the space \intersect memory region.
798 void CompactibleFreeListSpace::oop_iterate(MemRegion mr, OopClosure* cl) {
799 assert_lock_strong(freelistLock());
800 if (is_empty()) {
801 return;
802 }
803 MemRegion cur = MemRegion(bottom(), end());
804 mr = mr.intersection(cur);
805 if (mr.is_empty()) {
806 return;
807 }
808 if (mr.equals(cur)) {
809 oop_iterate(cl);
810 return;
811 }
812 assert(mr.end() <= end(), "just took an intersection above");
813 HeapWord* obj_addr = block_start(mr.start());
814 HeapWord* t = mr.end();
815
816 SpaceMemRegionOopsIterClosure smr_blk(cl, mr);
817 if (block_is_obj(obj_addr)) {
818 // Handle first object specially.
819 oop obj = oop(obj_addr);
820 obj_addr += adjustObjectSize(obj->oop_iterate(&smr_blk));
821 } else {
822 FreeChunk* fc = (FreeChunk*)obj_addr;
823 obj_addr += fc->size();
824 }
825 while (obj_addr < t) {
826 HeapWord* obj = obj_addr;
827 obj_addr += block_size(obj_addr);
828 // If "obj_addr" is not greater than top, then the
829 // entire object "obj" is within the region.
830 if (obj_addr <= t) {
831 if (block_is_obj(obj)) {
832 oop(obj)->oop_iterate(cl);
833 }
834 } else {
835 // "obj" extends beyond end of region
836 if (block_is_obj(obj)) {
837 oop(obj)->oop_iterate(&smr_blk);
838 }
839 break;
840 }
841 }
842 }
843
844 // NOTE: In the following methods, in order to safely be able to
845 // apply the closure to an object, we need to be sure that the
846 // object has been initialized. We are guaranteed that an object
847 // is initialized if we are holding the Heap_lock with the
848 // world stopped.
849 void CompactibleFreeListSpace::verify_objects_initialized() const {
850 if (is_init_completed()) {
851 assert_locked_or_safepoint(Heap_lock);
852 if (Universe::is_fully_initialized()) {
853 guarantee(SafepointSynchronize::is_at_safepoint(),
854 "Required for objects to be initialized");
855 }
856 } // else make a concession at vm start-up
857 }
858
859 // Apply the given closure to each object in the space
860 void CompactibleFreeListSpace::object_iterate(ObjectClosure* blk) {
861 assert_lock_strong(freelistLock());
862 NOT_PRODUCT(verify_objects_initialized());
863 HeapWord *cur, *limit;
864 size_t curSize;
865 for (cur = bottom(), limit = end(); cur < limit;
866 cur += curSize) {
867 curSize = block_size(cur);
868 if (block_is_obj(cur)) {
869 blk->do_object(oop(cur));
870 }
871 }
872 }
873
874 // Apply the given closure to each live object in the space
875 // The usage of CompactibleFreeListSpace
876 // by the ConcurrentMarkSweepGeneration for concurrent GC's allows
877 // objects in the space with references to objects that are no longer
878 // valid. For example, an object may reference another object
879 // that has already been sweep up (collected). This method uses
880 // obj_is_alive() to determine whether it is safe to apply the closure to
881 // an object. See obj_is_alive() for details on how liveness of an
882 // object is decided.
883
884 void CompactibleFreeListSpace::safe_object_iterate(ObjectClosure* blk) {
885 assert_lock_strong(freelistLock());
886 NOT_PRODUCT(verify_objects_initialized());
887 HeapWord *cur, *limit;
888 size_t curSize;
889 for (cur = bottom(), limit = end(); cur < limit;
890 cur += curSize) {
891 curSize = block_size(cur);
892 if (block_is_obj(cur) && obj_is_alive(cur)) {
893 blk->do_object(oop(cur));
894 }
895 }
896 }
897
898 void CompactibleFreeListSpace::object_iterate_mem(MemRegion mr,
899 UpwardsObjectClosure* cl) {
900 assert_locked(freelistLock());
901 NOT_PRODUCT(verify_objects_initialized());
902 Space::object_iterate_mem(mr, cl);
903 }
904
905 // Callers of this iterator beware: The closure application should
906 // be robust in the face of uninitialized objects and should (always)
907 // return a correct size so that the next addr + size below gives us a
908 // valid block boundary. [See for instance,
909 // ScanMarkedObjectsAgainCarefullyClosure::do_object_careful()
910 // in ConcurrentMarkSweepGeneration.cpp.]
911 HeapWord*
912 CompactibleFreeListSpace::object_iterate_careful(ObjectClosureCareful* cl) {
913 assert_lock_strong(freelistLock());
914 HeapWord *addr, *last;
915 size_t size;
916 for (addr = bottom(), last = end();
917 addr < last; addr += size) {
918 FreeChunk* fc = (FreeChunk*)addr;
919 if (fc->is_free()) {
920 // Since we hold the free list lock, which protects direct
921 // allocation in this generation by mutators, a free object
922 // will remain free throughout this iteration code.
923 size = fc->size();
924 } else {
925 // Note that the object need not necessarily be initialized,
926 // because (for instance) the free list lock does NOT protect
927 // object initialization. The closure application below must
928 // therefore be correct in the face of uninitialized objects.
929 size = cl->do_object_careful(oop(addr));
930 if (size == 0) {
931 // An unparsable object found. Signal early termination.
932 return addr;
933 }
934 }
935 }
936 return NULL;
937 }
938
939 // Callers of this iterator beware: The closure application should
940 // be robust in the face of uninitialized objects and should (always)
941 // return a correct size so that the next addr + size below gives us a
942 // valid block boundary. [See for instance,
943 // ScanMarkedObjectsAgainCarefullyClosure::do_object_careful()
944 // in ConcurrentMarkSweepGeneration.cpp.]
945 HeapWord*
946 CompactibleFreeListSpace::object_iterate_careful_m(MemRegion mr,
947 ObjectClosureCareful* cl) {
948 assert_lock_strong(freelistLock());
949 // Can't use used_region() below because it may not necessarily
950 // be the same as [bottom(),end()); although we could
951 // use [used_region().start(),round_to(used_region().end(),CardSize)),
952 // that appears too cumbersome, so we just do the simpler check
953 // in the assertion below.
954 assert(!mr.is_empty() && MemRegion(bottom(),end()).contains(mr),
955 "mr should be non-empty and within used space");
956 HeapWord *addr, *end;
957 size_t size;
958 for (addr = block_start_careful(mr.start()), end = mr.end();
959 addr < end; addr += size) {
960 FreeChunk* fc = (FreeChunk*)addr;
961 if (fc->is_free()) {
962 // Since we hold the free list lock, which protects direct
963 // allocation in this generation by mutators, a free object
964 // will remain free throughout this iteration code.
965 size = fc->size();
966 } else {
967 // Note that the object need not necessarily be initialized,
968 // because (for instance) the free list lock does NOT protect
969 // object initialization. The closure application below must
970 // therefore be correct in the face of uninitialized objects.
971 size = cl->do_object_careful_m(oop(addr), mr);
972 if (size == 0) {
973 // An unparsable object found. Signal early termination.
974 return addr;
975 }
976 }
977 }
978 return NULL;
979 }
980
981
982 HeapWord* CompactibleFreeListSpace::block_start_const(const void* p) const {
983 NOT_PRODUCT(verify_objects_initialized());
984 return _bt.block_start(p);
985 }
986
987 HeapWord* CompactibleFreeListSpace::block_start_careful(const void* p) const {
988 return _bt.block_start_careful(p);
989 }
990
991 size_t CompactibleFreeListSpace::block_size(const HeapWord* p) const {
992 NOT_PRODUCT(verify_objects_initialized());
993 // This must be volatile, or else there is a danger that the compiler
994 // will compile the code below into a sometimes-infinite loop, by keeping
995 // the value read the first time in a register.
996 while (true) {
997 // We must do this until we get a consistent view of the object.
998 if (FreeChunk::indicatesFreeChunk(p)) {
999 volatile FreeChunk* fc = (volatile FreeChunk*)p;
1000 size_t res = fc->size();
1001 // If the object is still a free chunk, return the size, else it
1002 // has been allocated so try again.
1003 if (FreeChunk::indicatesFreeChunk(p)) {
1004 assert(res != 0, "Block size should not be 0");
1005 return res;
1006 }
1007 } else {
1008 // must read from what 'p' points to in each loop.
1009 klassOop k = ((volatile oopDesc*)p)->klass_or_null();
1010 if (k != NULL) {
1011 assert(k->is_oop(true /* ignore mark word */), "Should be klass oop");
1012 oop o = (oop)p;
1013 assert(o->is_parsable(), "Should be parsable");
1014 assert(o->is_oop(true /* ignore mark word */), "Should be an oop.");
1015 size_t res = o->size_given_klass(k->klass_part());
1016 res = adjustObjectSize(res);
1017 assert(res != 0, "Block size should not be 0");
1018 return res;
1019 }
1020 }
1021 }
1022 }
1023
1024 // A variant of the above that uses the Printezis bits for
1025 // unparsable but allocated objects. This avoids any possible
1026 // stalls waiting for mutators to initialize objects, and is
1027 // thus potentially faster than the variant above. However,
1028 // this variant may return a zero size for a block that is
1029 // under mutation and for which a consistent size cannot be
1030 // inferred without stalling; see CMSCollector::block_size_if_printezis_bits().
1031 size_t CompactibleFreeListSpace::block_size_no_stall(HeapWord* p,
1032 const CMSCollector* c)
1033 const {
1034 assert(MemRegion(bottom(), end()).contains(p), "p not in space");
1035 // This must be volatile, or else there is a danger that the compiler
1036 // will compile the code below into a sometimes-infinite loop, by keeping
1037 // the value read the first time in a register.
1038 DEBUG_ONLY(uint loops = 0;)
1039 while (true) {
1040 // We must do this until we get a consistent view of the object.
1041 if (FreeChunk::indicatesFreeChunk(p)) {
1042 volatile FreeChunk* fc = (volatile FreeChunk*)p;
1043 size_t res = fc->size();
1044 if (FreeChunk::indicatesFreeChunk(p)) {
1045 assert(res != 0, "Block size should not be 0");
1046 assert(loops == 0, "Should be 0");
1047 return res;
1048 }
1049 } else {
1050 // must read from what 'p' points to in each loop.
1051 klassOop k = ((volatile oopDesc*)p)->klass_or_null();
1052 // We trust the size of any object that has a non-NULL
1053 // klass and (for those in the perm gen) is parsable
1054 // -- irrespective of its conc_safe-ty.
1055 if (k != NULL && ((oopDesc*)p)->is_parsable()) {
1056 assert(k->is_oop(), "Should really be klass oop.");
1057 oop o = (oop)p;
1058 assert(o->is_oop(), "Should be an oop");
1059 size_t res = o->size_given_klass(k->klass_part());
1060 res = adjustObjectSize(res);
1061 assert(res != 0, "Block size should not be 0");
1062 return res;
1063 } else {
1064 // May return 0 if P-bits not present.
1065 return c->block_size_if_printezis_bits(p);
1066 }
1067 }
1068 assert(loops == 0, "Can loop at most once");
1069 DEBUG_ONLY(loops++;)
1070 }
1071 }
1072
1073 size_t CompactibleFreeListSpace::block_size_nopar(const HeapWord* p) const {
1074 NOT_PRODUCT(verify_objects_initialized());
1075 assert(MemRegion(bottom(), end()).contains(p), "p not in space");
1076 FreeChunk* fc = (FreeChunk*)p;
1077 if (fc->is_free()) {
1078 return fc->size();
1079 } else {
1080 // Ignore mark word because this may be a recently promoted
1081 // object whose mark word is used to chain together grey
1082 // objects (the last one would have a null value).
1083 assert(oop(p)->is_oop(true), "Should be an oop");
1084 return adjustObjectSize(oop(p)->size());
1085 }
1086 }
1087
1088 // This implementation assumes that the property of "being an object" is
1089 // stable. But being a free chunk may not be (because of parallel
1090 // promotion.)
1091 bool CompactibleFreeListSpace::block_is_obj(const HeapWord* p) const {
1092 FreeChunk* fc = (FreeChunk*)p;
1093 assert(is_in_reserved(p), "Should be in space");
1094 // When doing a mark-sweep-compact of the CMS generation, this
1095 // assertion may fail because prepare_for_compaction() uses
1096 // space that is garbage to maintain information on ranges of
1097 // live objects so that these live ranges can be moved as a whole.
1098 // Comment out this assertion until that problem can be solved
1099 // (i.e., that the block start calculation may look at objects
1100 // at address below "p" in finding the object that contains "p"
1101 // and those objects (if garbage) may have been modified to hold
1102 // live range information.
1103 // assert(CollectedHeap::use_parallel_gc_threads() || _bt.block_start(p) == p,
1104 // "Should be a block boundary");
1105 if (FreeChunk::indicatesFreeChunk(p)) return false;
1106 klassOop k = oop(p)->klass_or_null();
1107 if (k != NULL) {
1108 // Ignore mark word because it may have been used to
1109 // chain together promoted objects (the last one
1110 // would have a null value).
1111 assert(oop(p)->is_oop(true), "Should be an oop");
1112 return true;
1113 } else {
1114 return false; // Was not an object at the start of collection.
1115 }
1116 }
1117
1118 // Check if the object is alive. This fact is checked either by consulting
1119 // the main marking bitmap in the sweeping phase or, if it's a permanent
1120 // generation and we're not in the sweeping phase, by checking the
1121 // perm_gen_verify_bit_map where we store the "deadness" information if
1122 // we did not sweep the perm gen in the most recent previous GC cycle.
1123 bool CompactibleFreeListSpace::obj_is_alive(const HeapWord* p) const {
1124 assert(SafepointSynchronize::is_at_safepoint() || !is_init_completed(),
1125 "Else races are possible");
1126 assert(block_is_obj(p), "The address should point to an object");
1127
1128 // If we're sweeping, we use object liveness information from the main bit map
1129 // for both perm gen and old gen.
1130 // We don't need to lock the bitmap (live_map or dead_map below), because
1131 // EITHER we are in the middle of the sweeping phase, and the
1132 // main marking bit map (live_map below) is locked,
1133 // OR we're in other phases and perm_gen_verify_bit_map (dead_map below)
1134 // is stable, because it's mutated only in the sweeping phase.
1135 // NOTE: This method is also used by jmap where, if class unloading is
1136 // off, the results can return "false" for legitimate perm objects,
1137 // when we are not in the midst of a sweeping phase, which can result
1138 // in jmap not reporting certain perm gen objects. This will be moot
1139 // if/when the perm gen goes away in the future.
1140 if (_collector->abstract_state() == CMSCollector::Sweeping) {
1141 CMSBitMap* live_map = _collector->markBitMap();
1142 return live_map->par_isMarked((HeapWord*) p);
1143 } else {
1144 // If we're not currently sweeping and we haven't swept the perm gen in
1145 // the previous concurrent cycle then we may have dead but unswept objects
1146 // in the perm gen. In this case, we use the "deadness" information
1147 // that we had saved in perm_gen_verify_bit_map at the last sweep.
1148 if (!CMSClassUnloadingEnabled && _collector->_permGen->reserved().contains(p)) {
1149 if (_collector->verifying()) {
1150 CMSBitMap* dead_map = _collector->perm_gen_verify_bit_map();
1151 // Object is marked in the dead_map bitmap at the previous sweep
1152 // when we know that it's dead; if the bitmap is not allocated then
1153 // the object is alive.
1154 return (dead_map->sizeInBits() == 0) // bit_map has been allocated
1155 || !dead_map->par_isMarked((HeapWord*) p);
1156 } else {
1157 return false; // We can't say for sure if it's live, so we say that it's dead.
1158 }
1159 }
1160 }
1161 return true;
1162 }
1163
1164 bool CompactibleFreeListSpace::block_is_obj_nopar(const HeapWord* p) const {
1165 FreeChunk* fc = (FreeChunk*)p;
1166 assert(is_in_reserved(p), "Should be in space");
1167 assert(_bt.block_start(p) == p, "Should be a block boundary");
1168 if (!fc->is_free()) {
1169 // Ignore mark word because it may have been used to
1170 // chain together promoted objects (the last one
1171 // would have a null value).
1172 assert(oop(p)->is_oop(true), "Should be an oop");
1173 return true;
1174 }
1175 return false;
1176 }
1177
1178 // "MT-safe but not guaranteed MT-precise" (TM); you may get an
1179 // approximate answer if you don't hold the freelistlock when you call this.
1180 size_t CompactibleFreeListSpace::totalSizeInIndexedFreeLists() const {
1181 size_t size = 0;
1182 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
1183 debug_only(
1184 // We may be calling here without the lock in which case we
1185 // won't do this modest sanity check.
1186 if (freelistLock()->owned_by_self()) {
1187 size_t total_list_size = 0;
1188 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL;
1189 fc = fc->next()) {
1190 total_list_size += i;
1191 }
1192 assert(total_list_size == i * _indexedFreeList[i].count(),
1193 "Count in list is incorrect");
1194 }
1195 )
1196 size += i * _indexedFreeList[i].count();
1197 }
1198 return size;
1199 }
1200
1201 HeapWord* CompactibleFreeListSpace::par_allocate(size_t size) {
1202 MutexLockerEx x(freelistLock(), Mutex::_no_safepoint_check_flag);
1203 return allocate(size);
1204 }
1205
1206 HeapWord*
1207 CompactibleFreeListSpace::getChunkFromSmallLinearAllocBlockRemainder(size_t size) {
1208 return getChunkFromLinearAllocBlockRemainder(&_smallLinearAllocBlock, size);
1209 }
1210
1211 HeapWord* CompactibleFreeListSpace::allocate(size_t size) {
1212 assert_lock_strong(freelistLock());
1213 HeapWord* res = NULL;
1214 assert(size == adjustObjectSize(size),
1215 "use adjustObjectSize() before calling into allocate()");
1216
1217 if (_adaptive_freelists) {
1218 res = allocate_adaptive_freelists(size);
1219 } else { // non-adaptive free lists
1220 res = allocate_non_adaptive_freelists(size);
1221 }
1222
1223 if (res != NULL) {
1224 // check that res does lie in this space!
1225 assert(is_in_reserved(res), "Not in this space!");
1226 assert(is_aligned((void*)res), "alignment check");
1227
1228 FreeChunk* fc = (FreeChunk*)res;
1229 fc->markNotFree();
1230 assert(!fc->is_free(), "shouldn't be marked free");
1231 assert(oop(fc)->klass_or_null() == NULL, "should look uninitialized");
1232 // Verify that the block offset table shows this to
1233 // be a single block, but not one which is unallocated.
1234 _bt.verify_single_block(res, size);
1235 _bt.verify_not_unallocated(res, size);
1236 // mangle a just allocated object with a distinct pattern.
1237 debug_only(fc->mangleAllocated(size));
1238 }
1239
1240 return res;
1241 }
1242
1243 HeapWord* CompactibleFreeListSpace::allocate_non_adaptive_freelists(size_t size) {
1244 HeapWord* res = NULL;
1245 // try and use linear allocation for smaller blocks
1246 if (size < _smallLinearAllocBlock._allocation_size_limit) {
1247 // if successful, the following also adjusts block offset table
1248 res = getChunkFromSmallLinearAllocBlock(size);
1249 }
1250 // Else triage to indexed lists for smaller sizes
1251 if (res == NULL) {
1252 if (size < SmallForDictionary) {
1253 res = (HeapWord*) getChunkFromIndexedFreeList(size);
1254 } else {
1255 // else get it from the big dictionary; if even this doesn't
1256 // work we are out of luck.
1257 res = (HeapWord*)getChunkFromDictionaryExact(size);
1258 }
1259 }
1260
1261 return res;
1262 }
1263
1264 HeapWord* CompactibleFreeListSpace::allocate_adaptive_freelists(size_t size) {
1265 assert_lock_strong(freelistLock());
1266 HeapWord* res = NULL;
1267 assert(size == adjustObjectSize(size),
1268 "use adjustObjectSize() before calling into allocate()");
1269
1270 // Strategy
1271 // if small
1272 // exact size from small object indexed list if small
1273 // small or large linear allocation block (linAB) as appropriate
1274 // take from lists of greater sized chunks
1275 // else
1276 // dictionary
1277 // small or large linear allocation block if it has the space
1278 // Try allocating exact size from indexTable first
1279 if (size < IndexSetSize) {
1280 res = (HeapWord*) getChunkFromIndexedFreeList(size);
1281 if(res != NULL) {
1282 assert(res != (HeapWord*)_indexedFreeList[size].head(),
1283 "Not removed from free list");
1284 // no block offset table adjustment is necessary on blocks in
1285 // the indexed lists.
1286
1287 // Try allocating from the small LinAB
1288 } else if (size < _smallLinearAllocBlock._allocation_size_limit &&
1289 (res = getChunkFromSmallLinearAllocBlock(size)) != NULL) {
1290 // if successful, the above also adjusts block offset table
1291 // Note that this call will refill the LinAB to
1292 // satisfy the request. This is different that
1293 // evm.
1294 // Don't record chunk off a LinAB? smallSplitBirth(size);
1295 } else {
1296 // Raid the exact free lists larger than size, even if they are not
1297 // overpopulated.
1298 res = (HeapWord*) getChunkFromGreater(size);
1299 }
1300 } else {
1301 // Big objects get allocated directly from the dictionary.
1302 res = (HeapWord*) getChunkFromDictionaryExact(size);
1303 if (res == NULL) {
1304 // Try hard not to fail since an allocation failure will likely
1305 // trigger a synchronous GC. Try to get the space from the
1306 // allocation blocks.
1307 res = getChunkFromSmallLinearAllocBlockRemainder(size);
1308 }
1309 }
1310
1311 return res;
1312 }
1313
1314 // A worst-case estimate of the space required (in HeapWords) to expand the heap
1315 // when promoting obj.
1316 size_t CompactibleFreeListSpace::expansionSpaceRequired(size_t obj_size) const {
1317 // Depending on the object size, expansion may require refilling either a
1318 // bigLAB or a smallLAB plus refilling a PromotionInfo object. MinChunkSize
1319 // is added because the dictionary may over-allocate to avoid fragmentation.
1320 size_t space = obj_size;
1321 if (!_adaptive_freelists) {
1322 space = MAX2(space, _smallLinearAllocBlock._refillSize);
1323 }
1324 space += _promoInfo.refillSize() + 2 * MinChunkSize;
1325 return space;
1326 }
1327
1328 FreeChunk* CompactibleFreeListSpace::getChunkFromGreater(size_t numWords) {
1329 FreeChunk* ret;
1330
1331 assert(numWords >= MinChunkSize, "Size is less than minimum");
1332 assert(linearAllocationWouldFail() || bestFitFirst(),
1333 "Should not be here");
1334
1335 size_t i;
1336 size_t currSize = numWords + MinChunkSize;
1337 assert(currSize % MinObjAlignment == 0, "currSize should be aligned");
1338 for (i = currSize; i < IndexSetSize; i += IndexSetStride) {
1339 FreeList<FreeChunk>* fl = &_indexedFreeList[i];
1340 if (fl->head()) {
1341 ret = getFromListGreater(fl, numWords);
1342 assert(ret == NULL || ret->is_free(), "Should be returning a free chunk");
1343 return ret;
1344 }
1345 }
1346
1347 currSize = MAX2((size_t)SmallForDictionary,
1348 (size_t)(numWords + MinChunkSize));
1349
1350 /* Try to get a chunk that satisfies request, while avoiding
1351 fragmentation that can't be handled. */
1352 {
1353 ret = dictionary()->get_chunk(currSize);
1354 if (ret != NULL) {
1355 assert(ret->size() - numWords >= MinChunkSize,
1356 "Chunk is too small");
1357 _bt.allocated((HeapWord*)ret, ret->size());
1358 /* Carve returned chunk. */
1359 (void) splitChunkAndReturnRemainder(ret, numWords);
1360 /* Label this as no longer a free chunk. */
1361 assert(ret->is_free(), "This chunk should be free");
1362 ret->link_prev(NULL);
1363 }
1364 assert(ret == NULL || ret->is_free(), "Should be returning a free chunk");
1365 return ret;
1366 }
1367 ShouldNotReachHere();
1368 }
1369
1370 bool CompactibleFreeListSpace::verifyChunkInIndexedFreeLists(FreeChunk* fc) const {
1371 assert(fc->size() < IndexSetSize, "Size of chunk is too large");
1372 return _indexedFreeList[fc->size()].verify_chunk_in_free_list(fc);
1373 }
1374
1375 bool CompactibleFreeListSpace::verify_chunk_is_linear_alloc_block(FreeChunk* fc) const {
1376 assert((_smallLinearAllocBlock._ptr != (HeapWord*)fc) ||
1377 (_smallLinearAllocBlock._word_size == fc->size()),
1378 "Linear allocation block shows incorrect size");
1379 return ((_smallLinearAllocBlock._ptr == (HeapWord*)fc) &&
1380 (_smallLinearAllocBlock._word_size == fc->size()));
1381 }
1382
1383 // Check if the purported free chunk is present either as a linear
1384 // allocation block, the size-indexed table of (smaller) free blocks,
1385 // or the larger free blocks kept in the binary tree dictionary.
1386 bool CompactibleFreeListSpace::verify_chunk_in_free_list(FreeChunk* fc) const {
1387 if (verify_chunk_is_linear_alloc_block(fc)) {
1388 return true;
1389 } else if (fc->size() < IndexSetSize) {
1390 return verifyChunkInIndexedFreeLists(fc);
1391 } else {
1392 return dictionary()->verify_chunk_in_free_list(fc);
1393 }
1394 }
1395
1396 #ifndef PRODUCT
1397 void CompactibleFreeListSpace::assert_locked() const {
1398 CMSLockVerifier::assert_locked(freelistLock(), parDictionaryAllocLock());
1399 }
1400
1401 void CompactibleFreeListSpace::assert_locked(const Mutex* lock) const {
1402 CMSLockVerifier::assert_locked(lock);
1403 }
1404 #endif
1405
1406 FreeChunk* CompactibleFreeListSpace::allocateScratch(size_t size) {
1407 // In the parallel case, the main thread holds the free list lock
1408 // on behalf the parallel threads.
1409 FreeChunk* fc;
1410 {
1411 // If GC is parallel, this might be called by several threads.
1412 // This should be rare enough that the locking overhead won't affect
1413 // the sequential code.
1414 MutexLockerEx x(parDictionaryAllocLock(),
1415 Mutex::_no_safepoint_check_flag);
1416 fc = getChunkFromDictionary(size);
1417 }
1418 if (fc != NULL) {
1419 fc->dontCoalesce();
1420 assert(fc->is_free(), "Should be free, but not coalescable");
1421 // Verify that the block offset table shows this to
1422 // be a single block, but not one which is unallocated.
1423 _bt.verify_single_block((HeapWord*)fc, fc->size());
1424 _bt.verify_not_unallocated((HeapWord*)fc, fc->size());
1425 }
1426 return fc;
1427 }
1428
1429 oop CompactibleFreeListSpace::promote(oop obj, size_t obj_size) {
1430 assert(obj_size == (size_t)obj->size(), "bad obj_size passed in");
1431 assert_locked();
1432
1433 // if we are tracking promotions, then first ensure space for
1434 // promotion (including spooling space for saving header if necessary).
1435 // then allocate and copy, then track promoted info if needed.
1436 // When tracking (see PromotionInfo::track()), the mark word may
1437 // be displaced and in this case restoration of the mark word
1438 // occurs in the (oop_since_save_marks_)iterate phase.
1439 if (_promoInfo.tracking() && !_promoInfo.ensure_spooling_space()) {
1440 return NULL;
1441 }
1442 // Call the allocate(size_t, bool) form directly to avoid the
1443 // additional call through the allocate(size_t) form. Having
1444 // the compile inline the call is problematic because allocate(size_t)
1445 // is a virtual method.
1446 HeapWord* res = allocate(adjustObjectSize(obj_size));
1447 if (res != NULL) {
1448 Copy::aligned_disjoint_words((HeapWord*)obj, res, obj_size);
1449 // if we should be tracking promotions, do so.
1450 if (_promoInfo.tracking()) {
1451 _promoInfo.track((PromotedObject*)res);
1452 }
1453 }
1454 return oop(res);
1455 }
1456
1457 HeapWord*
1458 CompactibleFreeListSpace::getChunkFromSmallLinearAllocBlock(size_t size) {
1459 assert_locked();
1460 assert(size >= MinChunkSize, "minimum chunk size");
1461 assert(size < _smallLinearAllocBlock._allocation_size_limit,
1462 "maximum from smallLinearAllocBlock");
1463 return getChunkFromLinearAllocBlock(&_smallLinearAllocBlock, size);
1464 }
1465
1466 HeapWord*
1467 CompactibleFreeListSpace::getChunkFromLinearAllocBlock(LinearAllocBlock *blk,
1468 size_t size) {
1469 assert_locked();
1470 assert(size >= MinChunkSize, "too small");
1471 HeapWord* res = NULL;
1472 // Try to do linear allocation from blk, making sure that
1473 if (blk->_word_size == 0) {
1474 // We have probably been unable to fill this either in the prologue or
1475 // when it was exhausted at the last linear allocation. Bail out until
1476 // next time.
1477 assert(blk->_ptr == NULL, "consistency check");
1478 return NULL;
1479 }
1480 assert(blk->_word_size != 0 && blk->_ptr != NULL, "consistency check");
1481 res = getChunkFromLinearAllocBlockRemainder(blk, size);
1482 if (res != NULL) return res;
1483
1484 // about to exhaust this linear allocation block
1485 if (blk->_word_size == size) { // exactly satisfied
1486 res = blk->_ptr;
1487 _bt.allocated(res, blk->_word_size);
1488 } else if (size + MinChunkSize <= blk->_refillSize) {
1489 size_t sz = blk->_word_size;
1490 // Update _unallocated_block if the size is such that chunk would be
1491 // returned to the indexed free list. All other chunks in the indexed
1492 // free lists are allocated from the dictionary so that _unallocated_block
1493 // has already been adjusted for them. Do it here so that the cost
1494 // for all chunks added back to the indexed free lists.
1495 if (sz < SmallForDictionary) {
1496 _bt.allocated(blk->_ptr, sz);
1497 }
1498 // Return the chunk that isn't big enough, and then refill below.
1499 addChunkToFreeLists(blk->_ptr, sz);
1500 split_birth(sz);
1501 // Don't keep statistics on adding back chunk from a LinAB.
1502 } else {
1503 // A refilled block would not satisfy the request.
1504 return NULL;
1505 }
1506
1507 blk->_ptr = NULL; blk->_word_size = 0;
1508 refillLinearAllocBlock(blk);
1509 assert(blk->_ptr == NULL || blk->_word_size >= size + MinChunkSize,
1510 "block was replenished");
1511 if (res != NULL) {
1512 split_birth(size);
1513 repairLinearAllocBlock(blk);
1514 } else if (blk->_ptr != NULL) {
1515 res = blk->_ptr;
1516 size_t blk_size = blk->_word_size;
1517 blk->_word_size -= size;
1518 blk->_ptr += size;
1519 split_birth(size);
1520 repairLinearAllocBlock(blk);
1521 // Update BOT last so that other (parallel) GC threads see a consistent
1522 // view of the BOT and free blocks.
1523 // Above must occur before BOT is updated below.
1524 OrderAccess::storestore();
1525 _bt.split_block(res, blk_size, size); // adjust block offset table
1526 }
1527 return res;
1528 }
1529
1530 HeapWord* CompactibleFreeListSpace::getChunkFromLinearAllocBlockRemainder(
1531 LinearAllocBlock* blk,
1532 size_t size) {
1533 assert_locked();
1534 assert(size >= MinChunkSize, "too small");
1535
1536 HeapWord* res = NULL;
1537 // This is the common case. Keep it simple.
1538 if (blk->_word_size >= size + MinChunkSize) {
1539 assert(blk->_ptr != NULL, "consistency check");
1540 res = blk->_ptr;
1541 // Note that the BOT is up-to-date for the linAB before allocation. It
1542 // indicates the start of the linAB. The split_block() updates the
1543 // BOT for the linAB after the allocation (indicates the start of the
1544 // next chunk to be allocated).
1545 size_t blk_size = blk->_word_size;
1546 blk->_word_size -= size;
1547 blk->_ptr += size;
1548 split_birth(size);
1549 repairLinearAllocBlock(blk);
1550 // Update BOT last so that other (parallel) GC threads see a consistent
1551 // view of the BOT and free blocks.
1552 // Above must occur before BOT is updated below.
1553 OrderAccess::storestore();
1554 _bt.split_block(res, blk_size, size); // adjust block offset table
1555 _bt.allocated(res, size);
1556 }
1557 return res;
1558 }
1559
1560 FreeChunk*
1561 CompactibleFreeListSpace::getChunkFromIndexedFreeList(size_t size) {
1562 assert_locked();
1563 assert(size < SmallForDictionary, "just checking");
1564 FreeChunk* res;
1565 res = _indexedFreeList[size].get_chunk_at_head();
1566 if (res == NULL) {
1567 res = getChunkFromIndexedFreeListHelper(size);
1568 }
1569 _bt.verify_not_unallocated((HeapWord*) res, size);
1570 assert(res == NULL || res->size() == size, "Incorrect block size");
1571 return res;
1572 }
1573
1574 FreeChunk*
1575 CompactibleFreeListSpace::getChunkFromIndexedFreeListHelper(size_t size,
1576 bool replenish) {
1577 assert_locked();
1578 FreeChunk* fc = NULL;
1579 if (size < SmallForDictionary) {
1580 assert(_indexedFreeList[size].head() == NULL ||
1581 _indexedFreeList[size].surplus() <= 0,
1582 "List for this size should be empty or under populated");
1583 // Try best fit in exact lists before replenishing the list
1584 if (!bestFitFirst() || (fc = bestFitSmall(size)) == NULL) {
1585 // Replenish list.
1586 //
1587 // Things tried that failed.
1588 // Tried allocating out of the two LinAB's first before
1589 // replenishing lists.
1590 // Tried small linAB of size 256 (size in indexed list)
1591 // and replenishing indexed lists from the small linAB.
1592 //
1593 FreeChunk* newFc = NULL;
1594 const size_t replenish_size = CMSIndexedFreeListReplenish * size;
1595 if (replenish_size < SmallForDictionary) {
1596 // Do not replenish from an underpopulated size.
1597 if (_indexedFreeList[replenish_size].surplus() > 0 &&
1598 _indexedFreeList[replenish_size].head() != NULL) {
1599 newFc = _indexedFreeList[replenish_size].get_chunk_at_head();
1600 } else if (bestFitFirst()) {
1601 newFc = bestFitSmall(replenish_size);
1602 }
1603 }
1604 if (newFc == NULL && replenish_size > size) {
1605 assert(CMSIndexedFreeListReplenish > 1, "ctl pt invariant");
1606 newFc = getChunkFromIndexedFreeListHelper(replenish_size, false);
1607 }
1608 // Note: The stats update re split-death of block obtained above
1609 // will be recorded below precisely when we know we are going to
1610 // be actually splitting it into more than one pieces below.
1611 if (newFc != NULL) {
1612 if (replenish || CMSReplenishIntermediate) {
1613 // Replenish this list and return one block to caller.
1614 size_t i;
1615 FreeChunk *curFc, *nextFc;
1616 size_t num_blk = newFc->size() / size;
1617 assert(num_blk >= 1, "Smaller than requested?");
1618 assert(newFc->size() % size == 0, "Should be integral multiple of request");
1619 if (num_blk > 1) {
1620 // we are sure we will be splitting the block just obtained
1621 // into multiple pieces; record the split-death of the original
1622 splitDeath(replenish_size);
1623 }
1624 // carve up and link blocks 0, ..., num_blk - 2
1625 // The last chunk is not added to the lists but is returned as the
1626 // free chunk.
1627 for (curFc = newFc, nextFc = (FreeChunk*)((HeapWord*)curFc + size),
1628 i = 0;
1629 i < (num_blk - 1);
1630 curFc = nextFc, nextFc = (FreeChunk*)((HeapWord*)nextFc + size),
1631 i++) {
1632 curFc->set_size(size);
1633 // Don't record this as a return in order to try and
1634 // determine the "returns" from a GC.
1635 _bt.verify_not_unallocated((HeapWord*) fc, size);
1636 _indexedFreeList[size].return_chunk_at_tail(curFc, false);
1637 _bt.mark_block((HeapWord*)curFc, size);
1638 split_birth(size);
1639 // Don't record the initial population of the indexed list
1640 // as a split birth.
1641 }
1642
1643 // check that the arithmetic was OK above
1644 assert((HeapWord*)nextFc == (HeapWord*)newFc + num_blk*size,
1645 "inconsistency in carving newFc");
1646 curFc->set_size(size);
1647 _bt.mark_block((HeapWord*)curFc, size);
1648 split_birth(size);
1649 fc = curFc;
1650 } else {
1651 // Return entire block to caller
1652 fc = newFc;
1653 }
1654 }
1655 }
1656 } else {
1657 // Get a free chunk from the free chunk dictionary to be returned to
1658 // replenish the indexed free list.
1659 fc = getChunkFromDictionaryExact(size);
1660 }
1661 // assert(fc == NULL || fc->is_free(), "Should be returning a free chunk");
1662 return fc;
1663 }
1664
1665 FreeChunk*
1666 CompactibleFreeListSpace::getChunkFromDictionary(size_t size) {
1667 assert_locked();
1668 FreeChunk* fc = _dictionary->get_chunk(size);
1669 if (fc == NULL) {
1670 return NULL;
1671 }
1672 _bt.allocated((HeapWord*)fc, fc->size());
1673 if (fc->size() >= size + MinChunkSize) {
1674 fc = splitChunkAndReturnRemainder(fc, size);
1675 }
1676 assert(fc->size() >= size, "chunk too small");
1677 assert(fc->size() < size + MinChunkSize, "chunk too big");
1678 _bt.verify_single_block((HeapWord*)fc, fc->size());
1679 return fc;
1680 }
1681
1682 FreeChunk*
1683 CompactibleFreeListSpace::getChunkFromDictionaryExact(size_t size) {
1684 assert_locked();
1685 FreeChunk* fc = _dictionary->get_chunk(size);
1686 if (fc == NULL) {
1687 return fc;
1688 }
1689 _bt.allocated((HeapWord*)fc, fc->size());
1690 if (fc->size() == size) {
1691 _bt.verify_single_block((HeapWord*)fc, size);
1692 return fc;
1693 }
1694 assert(fc->size() > size, "get_chunk() guarantee");
1695 if (fc->size() < size + MinChunkSize) {
1696 // Return the chunk to the dictionary and go get a bigger one.
1697 returnChunkToDictionary(fc);
1698 fc = _dictionary->get_chunk(size + MinChunkSize);
1699 if (fc == NULL) {
1700 return NULL;
1701 }
1702 _bt.allocated((HeapWord*)fc, fc->size());
1703 }
1704 assert(fc->size() >= size + MinChunkSize, "tautology");
1705 fc = splitChunkAndReturnRemainder(fc, size);
1706 assert(fc->size() == size, "chunk is wrong size");
1707 _bt.verify_single_block((HeapWord*)fc, size);
1708 return fc;
1709 }
1710
1711 void
1712 CompactibleFreeListSpace::returnChunkToDictionary(FreeChunk* chunk) {
1713 assert_locked();
1714
1715 size_t size = chunk->size();
1716 _bt.verify_single_block((HeapWord*)chunk, size);
1717 // adjust _unallocated_block downward, as necessary
1718 _bt.freed((HeapWord*)chunk, size);
1719 _dictionary->return_chunk(chunk);
1720 #ifndef PRODUCT
1721 if (CMSCollector::abstract_state() != CMSCollector::Sweeping) {
1722 TreeChunk<FreeChunk>::as_TreeChunk(chunk)->list()->verify_stats();
1723 }
1724 #endif // PRODUCT
1725 }
1726
1727 void
1728 CompactibleFreeListSpace::returnChunkToFreeList(FreeChunk* fc) {
1729 assert_locked();
1730 size_t size = fc->size();
1731 _bt.verify_single_block((HeapWord*) fc, size);
1732 _bt.verify_not_unallocated((HeapWord*) fc, size);
1733 if (_adaptive_freelists) {
1734 _indexedFreeList[size].return_chunk_at_tail(fc);
1735 } else {
1736 _indexedFreeList[size].return_chunk_at_head(fc);
1737 }
1738 #ifndef PRODUCT
1739 if (CMSCollector::abstract_state() != CMSCollector::Sweeping) {
1740 _indexedFreeList[size].verify_stats();
1741 }
1742 #endif // PRODUCT
1743 }
1744
1745 // Add chunk to end of last block -- if it's the largest
1746 // block -- and update BOT and census data. We would
1747 // of course have preferred to coalesce it with the
1748 // last block, but it's currently less expensive to find the
1749 // largest block than it is to find the last.
1750 void
1751 CompactibleFreeListSpace::addChunkToFreeListsAtEndRecordingStats(
1752 HeapWord* chunk, size_t size) {
1753 // check that the chunk does lie in this space!
1754 assert(chunk != NULL && is_in_reserved(chunk), "Not in this space!");
1755 // One of the parallel gc task threads may be here
1756 // whilst others are allocating.
1757 Mutex* lock = NULL;
1758 if (ParallelGCThreads != 0) {
1759 lock = &_parDictionaryAllocLock;
1760 }
1761 FreeChunk* ec;
1762 {
1763 MutexLockerEx x(lock, Mutex::_no_safepoint_check_flag);
1764 ec = dictionary()->find_largest_dict(); // get largest block
1765 if (ec != NULL && ec->end() == chunk) {
1766 // It's a coterminal block - we can coalesce.
1767 size_t old_size = ec->size();
1768 coalDeath(old_size);
1769 removeChunkFromDictionary(ec);
1770 size += old_size;
1771 } else {
1772 ec = (FreeChunk*)chunk;
1773 }
1774 }
1775 ec->set_size(size);
1776 debug_only(ec->mangleFreed(size));
1777 if (size < SmallForDictionary) {
1778 lock = _indexedFreeListParLocks[size];
1779 }
1780 MutexLockerEx x(lock, Mutex::_no_safepoint_check_flag);
1781 addChunkAndRepairOffsetTable((HeapWord*)ec, size, true);
1782 // record the birth under the lock since the recording involves
1783 // manipulation of the list on which the chunk lives and
1784 // if the chunk is allocated and is the last on the list,
1785 // the list can go away.
1786 coalBirth(size);
1787 }
1788
1789 void
1790 CompactibleFreeListSpace::addChunkToFreeLists(HeapWord* chunk,
1791 size_t size) {
1792 // check that the chunk does lie in this space!
1793 assert(chunk != NULL && is_in_reserved(chunk), "Not in this space!");
1794 assert_locked();
1795 _bt.verify_single_block(chunk, size);
1796
1797 FreeChunk* fc = (FreeChunk*) chunk;
1798 fc->set_size(size);
1799 debug_only(fc->mangleFreed(size));
1800 if (size < SmallForDictionary) {
1801 returnChunkToFreeList(fc);
1802 } else {
1803 returnChunkToDictionary(fc);
1804 }
1805 }
1806
1807 void
1808 CompactibleFreeListSpace::addChunkAndRepairOffsetTable(HeapWord* chunk,
1809 size_t size, bool coalesced) {
1810 assert_locked();
1811 assert(chunk != NULL, "null chunk");
1812 if (coalesced) {
1813 // repair BOT
1814 _bt.single_block(chunk, size);
1815 }
1816 addChunkToFreeLists(chunk, size);
1817 }
1818
1819 // We _must_ find the purported chunk on our free lists;
1820 // we assert if we don't.
1821 void
1822 CompactibleFreeListSpace::removeFreeChunkFromFreeLists(FreeChunk* fc) {
1823 size_t size = fc->size();
1824 assert_locked();
1825 debug_only(verifyFreeLists());
1826 if (size < SmallForDictionary) {
1827 removeChunkFromIndexedFreeList(fc);
1828 } else {
1829 removeChunkFromDictionary(fc);
1830 }
1831 _bt.verify_single_block((HeapWord*)fc, size);
1832 debug_only(verifyFreeLists());
1833 }
1834
1835 void
1836 CompactibleFreeListSpace::removeChunkFromDictionary(FreeChunk* fc) {
1837 size_t size = fc->size();
1838 assert_locked();
1839 assert(fc != NULL, "null chunk");
1840 _bt.verify_single_block((HeapWord*)fc, size);
1841 _dictionary->remove_chunk(fc);
1842 // adjust _unallocated_block upward, as necessary
1843 _bt.allocated((HeapWord*)fc, size);
1844 }
1845
1846 void
1847 CompactibleFreeListSpace::removeChunkFromIndexedFreeList(FreeChunk* fc) {
1848 assert_locked();
1849 size_t size = fc->size();
1850 _bt.verify_single_block((HeapWord*)fc, size);
1851 NOT_PRODUCT(
1852 if (FLSVerifyIndexTable) {
1853 verifyIndexedFreeList(size);
1854 }
1855 )
1856 _indexedFreeList[size].remove_chunk(fc);
1857 NOT_PRODUCT(
1858 if (FLSVerifyIndexTable) {
1859 verifyIndexedFreeList(size);
1860 }
1861 )
1862 }
1863
1864 FreeChunk* CompactibleFreeListSpace::bestFitSmall(size_t numWords) {
1865 /* A hint is the next larger size that has a surplus.
1866 Start search at a size large enough to guarantee that
1867 the excess is >= MIN_CHUNK. */
1868 size_t start = align_object_size(numWords + MinChunkSize);
1869 if (start < IndexSetSize) {
1870 FreeList<FreeChunk>* it = _indexedFreeList;
1871 size_t hint = _indexedFreeList[start].hint();
1872 while (hint < IndexSetSize) {
1873 assert(hint % MinObjAlignment == 0, "hint should be aligned");
1874 FreeList<FreeChunk> *fl = &_indexedFreeList[hint];
1875 if (fl->surplus() > 0 && fl->head() != NULL) {
1876 // Found a list with surplus, reset original hint
1877 // and split out a free chunk which is returned.
1878 _indexedFreeList[start].set_hint(hint);
1879 FreeChunk* res = getFromListGreater(fl, numWords);
1880 assert(res == NULL || res->is_free(),
1881 "Should be returning a free chunk");
1882 return res;
1883 }
1884 hint = fl->hint(); /* keep looking */
1885 }
1886 /* None found. */
1887 it[start].set_hint(IndexSetSize);
1888 }
1889 return NULL;
1890 }
1891
1892 /* Requires fl->size >= numWords + MinChunkSize */
1893 FreeChunk* CompactibleFreeListSpace::getFromListGreater(FreeList<FreeChunk>* fl,
1894 size_t numWords) {
1895 FreeChunk *curr = fl->head();
1896 size_t oldNumWords = curr->size();
1897 assert(numWords >= MinChunkSize, "Word size is too small");
1898 assert(curr != NULL, "List is empty");
1899 assert(oldNumWords >= numWords + MinChunkSize,
1900 "Size of chunks in the list is too small");
1901
1902 fl->remove_chunk(curr);
1903 // recorded indirectly by splitChunkAndReturnRemainder -
1904 // smallSplit(oldNumWords, numWords);
1905 FreeChunk* new_chunk = splitChunkAndReturnRemainder(curr, numWords);
1906 // Does anything have to be done for the remainder in terms of
1907 // fixing the card table?
1908 assert(new_chunk == NULL || new_chunk->is_free(),
1909 "Should be returning a free chunk");
1910 return new_chunk;
1911 }
1912
1913 FreeChunk*
1914 CompactibleFreeListSpace::splitChunkAndReturnRemainder(FreeChunk* chunk,
1915 size_t new_size) {
1916 assert_locked();
1917 size_t size = chunk->size();
1918 assert(size > new_size, "Split from a smaller block?");
1919 assert(is_aligned(chunk), "alignment problem");
1920 assert(size == adjustObjectSize(size), "alignment problem");
1921 size_t rem_size = size - new_size;
1922 assert(rem_size == adjustObjectSize(rem_size), "alignment problem");
1923 assert(rem_size >= MinChunkSize, "Free chunk smaller than minimum");
1924 FreeChunk* ffc = (FreeChunk*)((HeapWord*)chunk + new_size);
1925 assert(is_aligned(ffc), "alignment problem");
1926 ffc->set_size(rem_size);
1927 ffc->link_next(NULL);
1928 ffc->link_prev(NULL); // Mark as a free block for other (parallel) GC threads.
1929 // Above must occur before BOT is updated below.
1930 // adjust block offset table
1931 OrderAccess::storestore();
1932 assert(chunk->is_free() && ffc->is_free(), "Error");
1933 _bt.split_block((HeapWord*)chunk, chunk->size(), new_size);
1934 if (rem_size < SmallForDictionary) {
1935 bool is_par = (SharedHeap::heap()->n_par_threads() > 0);
1936 if (is_par) _indexedFreeListParLocks[rem_size]->lock();
1937 assert(!is_par ||
1938 (SharedHeap::heap()->n_par_threads() ==
1939 SharedHeap::heap()->workers()->active_workers()), "Mismatch");
1940 returnChunkToFreeList(ffc);
1941 split(size, rem_size);
1942 if (is_par) _indexedFreeListParLocks[rem_size]->unlock();
1943 } else {
1944 returnChunkToDictionary(ffc);
1945 split(size ,rem_size);
1946 }
1947 chunk->set_size(new_size);
1948 return chunk;
1949 }
1950
1951 void
1952 CompactibleFreeListSpace::sweep_completed() {
1953 // Now that space is probably plentiful, refill linear
1954 // allocation blocks as needed.
1955 refillLinearAllocBlocksIfNeeded();
1956 }
1957
1958 void
1959 CompactibleFreeListSpace::gc_prologue() {
1960 assert_locked();
1961 if (PrintFLSStatistics != 0) {
1962 gclog_or_tty->print("Before GC:\n");
1963 reportFreeListStatistics();
1964 }
1965 refillLinearAllocBlocksIfNeeded();
1966 }
1967
1968 void
1969 CompactibleFreeListSpace::gc_epilogue() {
1970 assert_locked();
1971 if (PrintGCDetails && Verbose && !_adaptive_freelists) {
1972 if (_smallLinearAllocBlock._word_size == 0)
1973 warning("CompactibleFreeListSpace(epilogue):: Linear allocation failure");
1974 }
1975 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency");
1976 _promoInfo.stopTrackingPromotions();
1977 repairLinearAllocationBlocks();
1978 // Print Space's stats
1979 if (PrintFLSStatistics != 0) {
1980 gclog_or_tty->print("After GC:\n");
1981 reportFreeListStatistics();
1982 }
1983 }
1984
1985 // Iteration support, mostly delegated from a CMS generation
1986
1987 void CompactibleFreeListSpace::save_marks() {
1988 assert(Thread::current()->is_VM_thread(),
1989 "Global variable should only be set when single-threaded");
1990 // Mark the "end" of the used space at the time of this call;
1991 // note, however, that promoted objects from this point
1992 // on are tracked in the _promoInfo below.
1993 set_saved_mark_word(unallocated_block());
1994 #ifdef ASSERT
1995 // Check the sanity of save_marks() etc.
1996 MemRegion ur = used_region();
1997 MemRegion urasm = used_region_at_save_marks();
1998 assert(ur.contains(urasm),
1999 err_msg(" Error at save_marks(): [" PTR_FORMAT "," PTR_FORMAT ")"
2000 " should contain [" PTR_FORMAT "," PTR_FORMAT ")",
2001 ur.start(), ur.end(), urasm.start(), urasm.end()));
2002 #endif
2003 // inform allocator that promotions should be tracked.
2004 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency");
2005 _promoInfo.startTrackingPromotions();
2006 }
2007
2008 bool CompactibleFreeListSpace::no_allocs_since_save_marks() {
2009 assert(_promoInfo.tracking(), "No preceding save_marks?");
2010 assert(SharedHeap::heap()->n_par_threads() == 0,
2011 "Shouldn't be called if using parallel gc.");
2012 return _promoInfo.noPromotions();
2013 }
2014
2015 #define CFLS_OOP_SINCE_SAVE_MARKS_DEFN(OopClosureType, nv_suffix) \
2016 \
2017 void CompactibleFreeListSpace:: \
2018 oop_since_save_marks_iterate##nv_suffix(OopClosureType* blk) { \
2019 assert(SharedHeap::heap()->n_par_threads() == 0, \
2020 "Shouldn't be called (yet) during parallel part of gc."); \
2021 _promoInfo.promoted_oops_iterate##nv_suffix(blk); \
2022 /* \
2023 * This also restores any displaced headers and removes the elements from \
2024 * the iteration set as they are processed, so that we have a clean slate \
2025 * at the end of the iteration. Note, thus, that if new objects are \
2026 * promoted as a result of the iteration they are iterated over as well. \
2027 */ \
2028 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency"); \
2029 }
2030
2031 ALL_SINCE_SAVE_MARKS_CLOSURES(CFLS_OOP_SINCE_SAVE_MARKS_DEFN)
2032
2033
2034 void CompactibleFreeListSpace::object_iterate_since_last_GC(ObjectClosure* cl) {
2035 // ugghh... how would one do this efficiently for a non-contiguous space?
2036 guarantee(false, "NYI");
2037 }
2038
2039 bool CompactibleFreeListSpace::linearAllocationWouldFail() const {
2040 return _smallLinearAllocBlock._word_size == 0;
2041 }
2042
2043 void CompactibleFreeListSpace::repairLinearAllocationBlocks() {
2044 // Fix up linear allocation blocks to look like free blocks
2045 repairLinearAllocBlock(&_smallLinearAllocBlock);
2046 }
2047
2048 void CompactibleFreeListSpace::repairLinearAllocBlock(LinearAllocBlock* blk) {
2049 assert_locked();
2050 if (blk->_ptr != NULL) {
2051 assert(blk->_word_size != 0 && blk->_word_size >= MinChunkSize,
2052 "Minimum block size requirement");
2053 FreeChunk* fc = (FreeChunk*)(blk->_ptr);
2054 fc->set_size(blk->_word_size);
2055 fc->link_prev(NULL); // mark as free
2056 fc->dontCoalesce();
2057 assert(fc->is_free(), "just marked it free");
2058 assert(fc->cantCoalesce(), "just marked it uncoalescable");
2059 }
2060 }
2061
2062 void CompactibleFreeListSpace::refillLinearAllocBlocksIfNeeded() {
2063 assert_locked();
2064 if (_smallLinearAllocBlock._ptr == NULL) {
2065 assert(_smallLinearAllocBlock._word_size == 0,
2066 "Size of linAB should be zero if the ptr is NULL");
2067 // Reset the linAB refill and allocation size limit.
2068 _smallLinearAllocBlock.set(0, 0, 1024*SmallForLinearAlloc, SmallForLinearAlloc);
2069 }
2070 refillLinearAllocBlockIfNeeded(&_smallLinearAllocBlock);
2071 }
2072
2073 void
2074 CompactibleFreeListSpace::refillLinearAllocBlockIfNeeded(LinearAllocBlock* blk) {
2075 assert_locked();
2076 assert((blk->_ptr == NULL && blk->_word_size == 0) ||
2077 (blk->_ptr != NULL && blk->_word_size >= MinChunkSize),
2078 "blk invariant");
2079 if (blk->_ptr == NULL) {
2080 refillLinearAllocBlock(blk);
2081 }
2082 if (PrintMiscellaneous && Verbose) {
2083 if (blk->_word_size == 0) {
2084 warning("CompactibleFreeListSpace(prologue):: Linear allocation failure");
2085 }
2086 }
2087 }
2088
2089 void
2090 CompactibleFreeListSpace::refillLinearAllocBlock(LinearAllocBlock* blk) {
2091 assert_locked();
2092 assert(blk->_word_size == 0 && blk->_ptr == NULL,
2093 "linear allocation block should be empty");
2094 FreeChunk* fc;
2095 if (blk->_refillSize < SmallForDictionary &&
2096 (fc = getChunkFromIndexedFreeList(blk->_refillSize)) != NULL) {
2097 // A linAB's strategy might be to use small sizes to reduce
2098 // fragmentation but still get the benefits of allocation from a
2099 // linAB.
2100 } else {
2101 fc = getChunkFromDictionary(blk->_refillSize);
2102 }
2103 if (fc != NULL) {
2104 blk->_ptr = (HeapWord*)fc;
2105 blk->_word_size = fc->size();
2106 fc->dontCoalesce(); // to prevent sweeper from sweeping us up
2107 }
2108 }
2109
2110 // Support for concurrent collection policy decisions.
2111 bool CompactibleFreeListSpace::should_concurrent_collect() const {
2112 // In the future we might want to add in frgamentation stats --
2113 // including erosion of the "mountain" into this decision as well.
2114 return !adaptive_freelists() && linearAllocationWouldFail();
2115 }
2116
2117 // Support for compaction
2118
2119 void CompactibleFreeListSpace::prepare_for_compaction(CompactPoint* cp) {
2120 SCAN_AND_FORWARD(cp,end,block_is_obj,block_size);
2121 // prepare_for_compaction() uses the space between live objects
2122 // so that later phase can skip dead space quickly. So verification
2123 // of the free lists doesn't work after.
2124 }
2125
2126 #define obj_size(q) adjustObjectSize(oop(q)->size())
2127 #define adjust_obj_size(s) adjustObjectSize(s)
2128
2129 void CompactibleFreeListSpace::adjust_pointers() {
2130 // In other versions of adjust_pointers(), a bail out
2131 // based on the amount of live data in the generation
2132 // (i.e., if 0, bail out) may be used.
2133 // Cannot test used() == 0 here because the free lists have already
2134 // been mangled by the compaction.
2135
2136 SCAN_AND_ADJUST_POINTERS(adjust_obj_size);
2137 // See note about verification in prepare_for_compaction().
2138 }
2139
2140 void CompactibleFreeListSpace::compact() {
2141 SCAN_AND_COMPACT(obj_size);
2142 }
2143
2144 // fragmentation_metric = 1 - [sum of (fbs**2) / (sum of fbs)**2]
2145 // where fbs is free block sizes
2146 double CompactibleFreeListSpace::flsFrag() const {
2147 size_t itabFree = totalSizeInIndexedFreeLists();
2148 double frag = 0.0;
2149 size_t i;
2150
2151 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
2152 double sz = i;
2153 frag += _indexedFreeList[i].count() * (sz * sz);
2154 }
2155
2156 double totFree = itabFree +
2157 _dictionary->total_chunk_size(DEBUG_ONLY(freelistLock()));
2158 if (totFree > 0) {
2159 frag = ((frag + _dictionary->sum_of_squared_block_sizes()) /
2160 (totFree * totFree));
2161 frag = (double)1.0 - frag;
2162 } else {
2163 assert(frag == 0.0, "Follows from totFree == 0");
2164 }
2165 return frag;
2166 }
2167
2168 void CompactibleFreeListSpace::beginSweepFLCensus(
2169 float inter_sweep_current,
2170 float inter_sweep_estimate,
2171 float intra_sweep_estimate) {
2172 assert_locked();
2173 size_t i;
2174 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
2175 FreeList<FreeChunk>* fl = &_indexedFreeList[i];
2176 if (PrintFLSStatistics > 1) {
2177 gclog_or_tty->print("size[%d] : ", i);
2178 }
2179 fl->compute_desired(inter_sweep_current, inter_sweep_estimate, intra_sweep_estimate);
2180 fl->set_coal_desired((ssize_t)((double)fl->desired() * CMSSmallCoalSurplusPercent));
2181 fl->set_before_sweep(fl->count());
2182 fl->set_bfr_surp(fl->surplus());
2183 }
2184 _dictionary->begin_sweep_dict_census(CMSLargeCoalSurplusPercent,
2185 inter_sweep_current,
2186 inter_sweep_estimate,
2187 intra_sweep_estimate);
2188 }
2189
2190 void CompactibleFreeListSpace::setFLSurplus() {
2191 assert_locked();
2192 size_t i;
2193 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
2194 FreeList<FreeChunk> *fl = &_indexedFreeList[i];
2195 fl->set_surplus(fl->count() -
2196 (ssize_t)((double)fl->desired() * CMSSmallSplitSurplusPercent));
2197 }
2198 }
2199
2200 void CompactibleFreeListSpace::setFLHints() {
2201 assert_locked();
2202 size_t i;
2203 size_t h = IndexSetSize;
2204 for (i = IndexSetSize - 1; i != 0; i -= IndexSetStride) {
2205 FreeList<FreeChunk> *fl = &_indexedFreeList[i];
2206 fl->set_hint(h);
2207 if (fl->surplus() > 0) {
2208 h = i;
2209 }
2210 }
2211 }
2212
2213 void CompactibleFreeListSpace::clearFLCensus() {
2214 assert_locked();
2215 size_t i;
2216 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
2217 FreeList<FreeChunk> *fl = &_indexedFreeList[i];
2218 fl->set_prev_sweep(fl->count());
2219 fl->set_coal_births(0);
2220 fl->set_coal_deaths(0);
2221 fl->set_split_births(0);
2222 fl->set_split_deaths(0);
2223 }
2224 }
2225
2226 void CompactibleFreeListSpace::endSweepFLCensus(size_t sweep_count) {
2227 if (PrintFLSStatistics > 0) {
2228 HeapWord* largestAddr = (HeapWord*) dictionary()->find_largest_dict();
2229 gclog_or_tty->print_cr("CMS: Large block " PTR_FORMAT,
2230 largestAddr);
2231 }
2232 setFLSurplus();
2233 setFLHints();
2234 if (PrintGC && PrintFLSCensus > 0) {
2235 printFLCensus(sweep_count);
2236 }
2237 clearFLCensus();
2238 assert_locked();
2239 _dictionary->end_sweep_dict_census(CMSLargeSplitSurplusPercent);
2240 }
2241
2242 bool CompactibleFreeListSpace::coalOverPopulated(size_t size) {
2243 if (size < SmallForDictionary) {
2244 FreeList<FreeChunk> *fl = &_indexedFreeList[size];
2245 return (fl->coal_desired() < 0) ||
2246 ((int)fl->count() > fl->coal_desired());
2247 } else {
2248 return dictionary()->coal_dict_over_populated(size);
2249 }
2250 }
2251
2252 void CompactibleFreeListSpace::smallCoalBirth(size_t size) {
2253 assert(size < SmallForDictionary, "Size too large for indexed list");
2254 FreeList<FreeChunk> *fl = &_indexedFreeList[size];
2255 fl->increment_coal_births();
2256 fl->increment_surplus();
2257 }
2258
2259 void CompactibleFreeListSpace::smallCoalDeath(size_t size) {
2260 assert(size < SmallForDictionary, "Size too large for indexed list");
2261 FreeList<FreeChunk> *fl = &_indexedFreeList[size];
2262 fl->increment_coal_deaths();
2263 fl->decrement_surplus();
2264 }
2265
2266 void CompactibleFreeListSpace::coalBirth(size_t size) {
2267 if (size < SmallForDictionary) {
2268 smallCoalBirth(size);
2269 } else {
2270 dictionary()->dict_census_udpate(size,
2271 false /* split */,
2272 true /* birth */);
2273 }
2274 }
2275
2276 void CompactibleFreeListSpace::coalDeath(size_t size) {
2277 if(size < SmallForDictionary) {
2278 smallCoalDeath(size);
2279 } else {
2280 dictionary()->dict_census_udpate(size,
2281 false /* split */,
2282 false /* birth */);
2283 }
2284 }
2285
2286 void CompactibleFreeListSpace::smallSplitBirth(size_t size) {
2287 assert(size < SmallForDictionary, "Size too large for indexed list");
2288 FreeList<FreeChunk> *fl = &_indexedFreeList[size];
2289 fl->increment_split_births();
2290 fl->increment_surplus();
2291 }
2292
2293 void CompactibleFreeListSpace::smallSplitDeath(size_t size) {
2294 assert(size < SmallForDictionary, "Size too large for indexed list");
2295 FreeList<FreeChunk> *fl = &_indexedFreeList[size];
2296 fl->increment_split_deaths();
2297 fl->decrement_surplus();
2298 }
2299
2300 void CompactibleFreeListSpace::split_birth(size_t size) {
2301 if (size < SmallForDictionary) {
2302 smallSplitBirth(size);
2303 } else {
2304 dictionary()->dict_census_udpate(size,
2305 true /* split */,
2306 true /* birth */);
2307 }
2308 }
2309
2310 void CompactibleFreeListSpace::splitDeath(size_t size) {
2311 if (size < SmallForDictionary) {
2312 smallSplitDeath(size);
2313 } else {
2314 dictionary()->dict_census_udpate(size,
2315 true /* split */,
2316 false /* birth */);
2317 }
2318 }
2319
2320 void CompactibleFreeListSpace::split(size_t from, size_t to1) {
2321 size_t to2 = from - to1;
2322 splitDeath(from);
2323 split_birth(to1);
2324 split_birth(to2);
2325 }
2326
2327 void CompactibleFreeListSpace::print() const {
2328 print_on(tty);
2329 }
2330
2331 void CompactibleFreeListSpace::prepare_for_verify() {
2332 assert_locked();
2333 repairLinearAllocationBlocks();
2334 // Verify that the SpoolBlocks look like free blocks of
2335 // appropriate sizes... To be done ...
2336 }
2337
2338 class VerifyAllBlksClosure: public BlkClosure {
2339 private:
2340 const CompactibleFreeListSpace* _sp;
2341 const MemRegion _span;
2342 HeapWord* _last_addr;
2343 size_t _last_size;
2344 bool _last_was_obj;
2345 bool _last_was_live;
2346
2347 public:
2348 VerifyAllBlksClosure(const CompactibleFreeListSpace* sp,
2349 MemRegion span) : _sp(sp), _span(span),
2350 _last_addr(NULL), _last_size(0),
2351 _last_was_obj(false), _last_was_live(false) { }
2352
2353 virtual size_t do_blk(HeapWord* addr) {
2354 size_t res;
2355 bool was_obj = false;
2356 bool was_live = false;
2357 if (_sp->block_is_obj(addr)) {
2358 was_obj = true;
2359 oop p = oop(addr);
2360 guarantee(p->is_oop(), "Should be an oop");
2361 res = _sp->adjustObjectSize(p->size());
2362 if (_sp->obj_is_alive(addr)) {
2363 was_live = true;
2364 p->verify();
2365 }
2366 } else {
2367 FreeChunk* fc = (FreeChunk*)addr;
2368 res = fc->size();
2369 if (FLSVerifyLists && !fc->cantCoalesce()) {
2370 guarantee(_sp->verify_chunk_in_free_list(fc),
2371 "Chunk should be on a free list");
2372 }
2373 }
2374 if (res == 0) {
2375 gclog_or_tty->print_cr("Livelock: no rank reduction!");
2376 gclog_or_tty->print_cr(
2377 " Current: addr = " PTR_FORMAT ", size = " SIZE_FORMAT ", obj = %s, live = %s \n"
2378 " Previous: addr = " PTR_FORMAT ", size = " SIZE_FORMAT ", obj = %s, live = %s \n",
2379 addr, res, was_obj ?"true":"false", was_live ?"true":"false",
2380 _last_addr, _last_size, _last_was_obj?"true":"false", _last_was_live?"true":"false");
2381 _sp->print_on(gclog_or_tty);
2382 guarantee(false, "Seppuku!");
2383 }
2384 _last_addr = addr;
2385 _last_size = res;
2386 _last_was_obj = was_obj;
2387 _last_was_live = was_live;
2388 return res;
2389 }
2390 };
2391
2392 class VerifyAllOopsClosure: public OopClosure {
2393 private:
2394 const CMSCollector* _collector;
2395 const CompactibleFreeListSpace* _sp;
2396 const MemRegion _span;
2397 const bool _past_remark;
2398 const CMSBitMap* _bit_map;
2399
2400 protected:
2401 void do_oop(void* p, oop obj) {
2402 if (_span.contains(obj)) { // the interior oop points into CMS heap
2403 if (!_span.contains(p)) { // reference from outside CMS heap
2404 // Should be a valid object; the first disjunct below allows
2405 // us to sidestep an assertion in block_is_obj() that insists
2406 // that p be in _sp. Note that several generations (and spaces)
2407 // are spanned by _span (CMS heap) above.
2408 guarantee(!_sp->is_in_reserved(obj) ||
2409 _sp->block_is_obj((HeapWord*)obj),
2410 "Should be an object");
2411 guarantee(obj->is_oop(), "Should be an oop");
2412 obj->verify();
2413 if (_past_remark) {
2414 // Remark has been completed, the object should be marked
2415 _bit_map->isMarked((HeapWord*)obj);
2416 }
2417 } else { // reference within CMS heap
2418 if (_past_remark) {
2419 // Remark has been completed -- so the referent should have
2420 // been marked, if referring object is.
2421 if (_bit_map->isMarked(_collector->block_start(p))) {
2422 guarantee(_bit_map->isMarked((HeapWord*)obj), "Marking error?");
2423 }
2424 }
2425 }
2426 } else if (_sp->is_in_reserved(p)) {
2427 // the reference is from FLS, and points out of FLS
2428 guarantee(obj->is_oop(), "Should be an oop");
2429 obj->verify();
2430 }
2431 }
2432
2433 template <class T> void do_oop_work(T* p) {
2434 T heap_oop = oopDesc::load_heap_oop(p);
2435 if (!oopDesc::is_null(heap_oop)) {
2436 oop obj = oopDesc::decode_heap_oop_not_null(heap_oop);
2437 do_oop(p, obj);
2438 }
2439 }
2440
2441 public:
2442 VerifyAllOopsClosure(const CMSCollector* collector,
2443 const CompactibleFreeListSpace* sp, MemRegion span,
2444 bool past_remark, CMSBitMap* bit_map) :
2445 OopClosure(), _collector(collector), _sp(sp), _span(span),
2446 _past_remark(past_remark), _bit_map(bit_map) { }
2447
2448 virtual void do_oop(oop* p) { VerifyAllOopsClosure::do_oop_work(p); }
2449 virtual void do_oop(narrowOop* p) { VerifyAllOopsClosure::do_oop_work(p); }
2450 };
2451
2452 void CompactibleFreeListSpace::verify() const {
2453 assert_lock_strong(&_freelistLock);
2454 verify_objects_initialized();
2455 MemRegion span = _collector->_span;
2456 bool past_remark = (_collector->abstract_state() ==
2457 CMSCollector::Sweeping);
2458
2459 ResourceMark rm;
2460 HandleMark hm;
2461
2462 // Check integrity of CFL data structures
2463 _promoInfo.verify();
2464 _dictionary->verify();
2465 if (FLSVerifyIndexTable) {
2466 verifyIndexedFreeLists();
2467 }
2468 // Check integrity of all objects and free blocks in space
2469 {
2470 VerifyAllBlksClosure cl(this, span);
2471 ((CompactibleFreeListSpace*)this)->blk_iterate(&cl); // cast off const
2472 }
2473 // Check that all references in the heap to FLS
2474 // are to valid objects in FLS or that references in
2475 // FLS are to valid objects elsewhere in the heap
2476 if (FLSVerifyAllHeapReferences)
2477 {
2478 VerifyAllOopsClosure cl(_collector, this, span, past_remark,
2479 _collector->markBitMap());
2480 CollectedHeap* ch = Universe::heap();
2481 ch->oop_iterate(&cl); // all oops in generations
2482 ch->permanent_oop_iterate(&cl); // all oops in perm gen
2483 }
2484
2485 if (VerifyObjectStartArray) {
2486 // Verify the block offset table
2487 _bt.verify();
2488 }
2489 }
2490
2491 #ifndef PRODUCT
2492 void CompactibleFreeListSpace::verifyFreeLists() const {
2493 if (FLSVerifyLists) {
2494 _dictionary->verify();
2495 verifyIndexedFreeLists();
2496 } else {
2497 if (FLSVerifyDictionary) {
2498 _dictionary->verify();
2499 }
2500 if (FLSVerifyIndexTable) {
2501 verifyIndexedFreeLists();
2502 }
2503 }
2504 }
2505 #endif
2506
2507 void CompactibleFreeListSpace::verifyIndexedFreeLists() const {
2508 size_t i = 0;
2509 for (; i < IndexSetStart; i++) {
2510 guarantee(_indexedFreeList[i].head() == NULL, "should be NULL");
2511 }
2512 for (; i < IndexSetSize; i++) {
2513 verifyIndexedFreeList(i);
2514 }
2515 }
2516
2517 void CompactibleFreeListSpace::verifyIndexedFreeList(size_t size) const {
2518 FreeChunk* fc = _indexedFreeList[size].head();
2519 FreeChunk* tail = _indexedFreeList[size].tail();
2520 size_t num = _indexedFreeList[size].count();
2521 size_t n = 0;
2522 guarantee(((size >= IndexSetStart) && (size % IndexSetStride == 0)) || fc == NULL,
2523 "Slot should have been empty");
2524 for (; fc != NULL; fc = fc->next(), n++) {
2525 guarantee(fc->size() == size, "Size inconsistency");
2526 guarantee(fc->is_free(), "!free?");
2527 guarantee(fc->next() == NULL || fc->next()->prev() == fc, "Broken list");
2528 guarantee((fc->next() == NULL) == (fc == tail), "Incorrect tail");
2529 }
2530 guarantee(n == num, "Incorrect count");
2531 }
2532
2533 #ifndef PRODUCT
2534 void CompactibleFreeListSpace::check_free_list_consistency() const {
2535 assert(_dictionary->min_size() <= IndexSetSize,
2536 "Some sizes can't be allocated without recourse to"
2537 " linear allocation buffers");
2538 assert(BinaryTreeDictionary<FreeChunk>::min_tree_chunk_size*HeapWordSize == sizeof(TreeChunk<FreeChunk>),
2539 "else MIN_TREE_CHUNK_SIZE is wrong");
2540 assert(IndexSetStart != 0, "IndexSetStart not initialized");
2541 assert(IndexSetStride != 0, "IndexSetStride not initialized");
2542 }
2543 #endif
2544
2545 void CompactibleFreeListSpace::printFLCensus(size_t sweep_count) const {
2546 assert_lock_strong(&_freelistLock);
2547 FreeList<FreeChunk> total;
2548 gclog_or_tty->print("end sweep# " SIZE_FORMAT "\n", sweep_count);
2549 FreeList<FreeChunk>::print_labels_on(gclog_or_tty, "size");
2550 size_t total_free = 0;
2551 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) {
2552 const FreeList<FreeChunk> *fl = &_indexedFreeList[i];
2553 total_free += fl->count() * fl->size();
2554 if (i % (40*IndexSetStride) == 0) {
2555 FreeList<FreeChunk>::print_labels_on(gclog_or_tty, "size");
2556 }
2557 fl->print_on(gclog_or_tty);
2558 total.set_bfr_surp( total.bfr_surp() + fl->bfr_surp() );
2559 total.set_surplus( total.surplus() + fl->surplus() );
2560 total.set_desired( total.desired() + fl->desired() );
2561 total.set_prev_sweep( total.prev_sweep() + fl->prev_sweep() );
2562 total.set_before_sweep(total.before_sweep() + fl->before_sweep());
2563 total.set_count( total.count() + fl->count() );
2564 total.set_coal_births( total.coal_births() + fl->coal_births() );
2565 total.set_coal_deaths( total.coal_deaths() + fl->coal_deaths() );
2566 total.set_split_births(total.split_births() + fl->split_births());
2567 total.set_split_deaths(total.split_deaths() + fl->split_deaths());
2568 }
2569 total.print_on(gclog_or_tty, "TOTAL");
2570 gclog_or_tty->print_cr("Total free in indexed lists "
2571 SIZE_FORMAT " words", total_free);
2572 gclog_or_tty->print("growth: %8.5f deficit: %8.5f\n",
2573 (double)(total.split_births()+total.coal_births()-total.split_deaths()-total.coal_deaths())/
2574 (total.prev_sweep() != 0 ? (double)total.prev_sweep() : 1.0),
2575 (double)(total.desired() - total.count())/(total.desired() != 0 ? (double)total.desired() : 1.0));
2576 _dictionary->print_dict_census();
2577 }
2578
2579 ///////////////////////////////////////////////////////////////////////////
2580 // CFLS_LAB
2581 ///////////////////////////////////////////////////////////////////////////
2582
2583 #define VECTOR_257(x) \
2584 /* 1 2 3 4 5 6 7 8 9 1x 11 12 13 14 15 16 17 18 19 2x 21 22 23 24 25 26 27 28 29 3x 31 32 */ \
2585 { x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2586 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2587 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2588 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2589 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2590 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2591 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2592 x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, \
2593 x }
2594
2595 // Initialize with default setting of CMSParPromoteBlocksToClaim, _not_
2596 // OldPLABSize, whose static default is different; if overridden at the
2597 // command-line, this will get reinitialized via a call to
2598 // modify_initialization() below.
2599 AdaptiveWeightedAverage CFLS_LAB::_blocks_to_claim[] =
2600 VECTOR_257(AdaptiveWeightedAverage(OldPLABWeight, (float)CMSParPromoteBlocksToClaim));
2601 size_t CFLS_LAB::_global_num_blocks[] = VECTOR_257(0);
2602 uint CFLS_LAB::_global_num_workers[] = VECTOR_257(0);
2603
2604 CFLS_LAB::CFLS_LAB(CompactibleFreeListSpace* cfls) :
2605 _cfls(cfls)
2606 {
2607 assert(CompactibleFreeListSpace::IndexSetSize == 257, "Modify VECTOR_257() macro above");
2608 for (size_t i = CompactibleFreeListSpace::IndexSetStart;
2609 i < CompactibleFreeListSpace::IndexSetSize;
2610 i += CompactibleFreeListSpace::IndexSetStride) {
2611 _indexedFreeList[i].set_size(i);
2612 _num_blocks[i] = 0;
2613 }
2614 }
2615
2616 static bool _CFLS_LAB_modified = false;
2617
2618 void CFLS_LAB::modify_initialization(size_t n, unsigned wt) {
2619 assert(!_CFLS_LAB_modified, "Call only once");
2620 _CFLS_LAB_modified = true;
2621 for (size_t i = CompactibleFreeListSpace::IndexSetStart;
2622 i < CompactibleFreeListSpace::IndexSetSize;
2623 i += CompactibleFreeListSpace::IndexSetStride) {
2624 _blocks_to_claim[i].modify(n, wt, true /* force */);
2625 }
2626 }
2627
2628 HeapWord* CFLS_LAB::alloc(size_t word_sz) {
2629 FreeChunk* res;
2630 assert(word_sz == _cfls->adjustObjectSize(word_sz), "Error");
2631 if (word_sz >= CompactibleFreeListSpace::IndexSetSize) {
2632 // This locking manages sync with other large object allocations.
2633 MutexLockerEx x(_cfls->parDictionaryAllocLock(),
2634 Mutex::_no_safepoint_check_flag);
2635 res = _cfls->getChunkFromDictionaryExact(word_sz);
2636 if (res == NULL) return NULL;
2637 } else {
2638 FreeList<FreeChunk>* fl = &_indexedFreeList[word_sz];
2639 if (fl->count() == 0) {
2640 // Attempt to refill this local free list.
2641 get_from_global_pool(word_sz, fl);
2642 // If it didn't work, give up.
2643 if (fl->count() == 0) return NULL;
2644 }
2645 res = fl->get_chunk_at_head();
2646 assert(res != NULL, "Why was count non-zero?");
2647 }
2648 res->markNotFree();
2649 assert(!res->is_free(), "shouldn't be marked free");
2650 assert(oop(res)->klass_or_null() == NULL, "should look uninitialized");
2651 // mangle a just allocated object with a distinct pattern.
2652 debug_only(res->mangleAllocated(word_sz));
2653 return (HeapWord*)res;
2654 }
2655
2656 // Get a chunk of blocks of the right size and update related
2657 // book-keeping stats
2658 void CFLS_LAB::get_from_global_pool(size_t word_sz, FreeList<FreeChunk>* fl) {
2659 // Get the #blocks we want to claim
2660 size_t n_blks = (size_t)_blocks_to_claim[word_sz].average();
2661 assert(n_blks > 0, "Error");
2662 assert(ResizePLAB || n_blks == OldPLABSize, "Error");
2663 // In some cases, when the application has a phase change,
2664 // there may be a sudden and sharp shift in the object survival
2665 // profile, and updating the counts at the end of a scavenge
2666 // may not be quick enough, giving rise to large scavenge pauses
2667 // during these phase changes. It is beneficial to detect such
2668 // changes on-the-fly during a scavenge and avoid such a phase-change
2669 // pothole. The following code is a heuristic attempt to do that.
2670 // It is protected by a product flag until we have gained
2671 // enough experience with this heuristic and fine-tuned its behaviour.
2672 // WARNING: This might increase fragmentation if we overreact to
2673 // small spikes, so some kind of historical smoothing based on
2674 // previous experience with the greater reactivity might be useful.
2675 // Lacking sufficient experience, CMSOldPLABResizeQuicker is disabled by
2676 // default.
2677 if (ResizeOldPLAB && CMSOldPLABResizeQuicker) {
2678 size_t multiple = _num_blocks[word_sz]/(CMSOldPLABToleranceFactor*CMSOldPLABNumRefills*n_blks);
2679 n_blks += CMSOldPLABReactivityFactor*multiple*n_blks;
2680 n_blks = MIN2(n_blks, CMSOldPLABMax);
2681 }
2682 assert(n_blks > 0, "Error");
2683 _cfls->par_get_chunk_of_blocks(word_sz, n_blks, fl);
2684 // Update stats table entry for this block size
2685 _num_blocks[word_sz] += fl->count();
2686 }
2687
2688 void CFLS_LAB::compute_desired_plab_size() {
2689 for (size_t i = CompactibleFreeListSpace::IndexSetStart;
2690 i < CompactibleFreeListSpace::IndexSetSize;
2691 i += CompactibleFreeListSpace::IndexSetStride) {
2692 assert((_global_num_workers[i] == 0) == (_global_num_blocks[i] == 0),
2693 "Counter inconsistency");
2694 if (_global_num_workers[i] > 0) {
2695 // Need to smooth wrt historical average
2696 if (ResizeOldPLAB) {
2697 _blocks_to_claim[i].sample(
2698 MAX2((size_t)CMSOldPLABMin,
2699 MIN2((size_t)CMSOldPLABMax,
2700 _global_num_blocks[i]/(_global_num_workers[i]*CMSOldPLABNumRefills))));
2701 }
2702 // Reset counters for next round
2703 _global_num_workers[i] = 0;
2704 _global_num_blocks[i] = 0;
2705 if (PrintOldPLAB) {
2706 gclog_or_tty->print_cr("[%d]: %d", i, (size_t)_blocks_to_claim[i].average());
2707 }
2708 }
2709 }
2710 }
2711
2712 // If this is changed in the future to allow parallel
2713 // access, one would need to take the FL locks and,
2714 // depending on how it is used, stagger access from
2715 // parallel threads to reduce contention.
2716 void CFLS_LAB::retire(int tid) {
2717 // We run this single threaded with the world stopped;
2718 // so no need for locks and such.
2719 NOT_PRODUCT(Thread* t = Thread::current();)
2720 assert(Thread::current()->is_VM_thread(), "Error");
2721 for (size_t i = CompactibleFreeListSpace::IndexSetStart;
2722 i < CompactibleFreeListSpace::IndexSetSize;
2723 i += CompactibleFreeListSpace::IndexSetStride) {
2724 assert(_num_blocks[i] >= (size_t)_indexedFreeList[i].count(),
2725 "Can't retire more than what we obtained");
2726 if (_num_blocks[i] > 0) {
2727 size_t num_retire = _indexedFreeList[i].count();
2728 assert(_num_blocks[i] > num_retire, "Should have used at least one");
2729 {
2730 // MutexLockerEx x(_cfls->_indexedFreeListParLocks[i],
2731 // Mutex::_no_safepoint_check_flag);
2732
2733 // Update globals stats for num_blocks used
2734 _global_num_blocks[i] += (_num_blocks[i] - num_retire);
2735 _global_num_workers[i]++;
2736 assert(_global_num_workers[i] <= ParallelGCThreads, "Too big");
2737 if (num_retire > 0) {
2738 _cfls->_indexedFreeList[i].prepend(&_indexedFreeList[i]);
2739 // Reset this list.
2740 _indexedFreeList[i] = FreeList<FreeChunk>();
2741 _indexedFreeList[i].set_size(i);
2742 }
2743 }
2744 if (PrintOldPLAB) {
2745 gclog_or_tty->print_cr("%d[%d]: %d/%d/%d",
2746 tid, i, num_retire, _num_blocks[i], (size_t)_blocks_to_claim[i].average());
2747 }
2748 // Reset stats for next round
2749 _num_blocks[i] = 0;
2750 }
2751 }
2752 }
2753
2754 void CompactibleFreeListSpace:: par_get_chunk_of_blocks(size_t word_sz, size_t n, FreeList<FreeChunk>* fl) {
2755 assert(fl->count() == 0, "Precondition.");
2756 assert(word_sz < CompactibleFreeListSpace::IndexSetSize,
2757 "Precondition");
2758
2759 // We'll try all multiples of word_sz in the indexed set, starting with
2760 // word_sz itself and, if CMSSplitIndexedFreeListBlocks, try larger multiples,
2761 // then try getting a big chunk and splitting it.
2762 {
2763 bool found;
2764 int k;
2765 size_t cur_sz;
2766 for (k = 1, cur_sz = k * word_sz, found = false;
2767 (cur_sz < CompactibleFreeListSpace::IndexSetSize) &&
2768 (CMSSplitIndexedFreeListBlocks || k <= 1);
2769 k++, cur_sz = k * word_sz) {
2770 FreeList<FreeChunk> fl_for_cur_sz; // Empty.
2771 fl_for_cur_sz.set_size(cur_sz);
2772 {
2773 MutexLockerEx x(_indexedFreeListParLocks[cur_sz],
2774 Mutex::_no_safepoint_check_flag);
2775 FreeList<FreeChunk>* gfl = &_indexedFreeList[cur_sz];
2776 if (gfl->count() != 0) {
2777 // nn is the number of chunks of size cur_sz that
2778 // we'd need to split k-ways each, in order to create
2779 // "n" chunks of size word_sz each.
2780 const size_t nn = MAX2(n/k, (size_t)1);
2781 gfl->getFirstNChunksFromList(nn, &fl_for_cur_sz);
2782 found = true;
2783 if (k > 1) {
2784 // Update split death stats for the cur_sz-size blocks list:
2785 // we increment the split death count by the number of blocks
2786 // we just took from the cur_sz-size blocks list and which
2787 // we will be splitting below.
2788 ssize_t deaths = gfl->split_deaths() +
2789 fl_for_cur_sz.count();
2790 gfl->set_split_deaths(deaths);
2791 }
2792 }
2793 }
2794 // Now transfer fl_for_cur_sz to fl. Common case, we hope, is k = 1.
2795 if (found) {
2796 if (k == 1) {
2797 fl->prepend(&fl_for_cur_sz);
2798 } else {
2799 // Divide each block on fl_for_cur_sz up k ways.
2800 FreeChunk* fc;
2801 while ((fc = fl_for_cur_sz.get_chunk_at_head()) != NULL) {
2802 // Must do this in reverse order, so that anybody attempting to
2803 // access the main chunk sees it as a single free block until we
2804 // change it.
2805 size_t fc_size = fc->size();
2806 assert(fc->is_free(), "Error");
2807 for (int i = k-1; i >= 0; i--) {
2808 FreeChunk* ffc = (FreeChunk*)((HeapWord*)fc + i * word_sz);
2809 assert((i != 0) ||
2810 ((fc == ffc) && ffc->is_free() &&
2811 (ffc->size() == k*word_sz) && (fc_size == word_sz)),
2812 "Counting error");
2813 ffc->set_size(word_sz);
2814 ffc->link_prev(NULL); // Mark as a free block for other (parallel) GC threads.
2815 ffc->link_next(NULL);
2816 // Above must occur before BOT is updated below.
2817 OrderAccess::storestore();
2818 // splitting from the right, fc_size == i * word_sz
2819 _bt.mark_block((HeapWord*)ffc, word_sz, true /* reducing */);
2820 fc_size -= word_sz;
2821 assert(fc_size == i*word_sz, "Error");
2822 _bt.verify_not_unallocated((HeapWord*)ffc, word_sz);
2823 _bt.verify_single_block((HeapWord*)fc, fc_size);
2824 _bt.verify_single_block((HeapWord*)ffc, word_sz);
2825 // Push this on "fl".
2826 fl->return_chunk_at_head(ffc);
2827 }
2828 // TRAP
2829 assert(fl->tail()->next() == NULL, "List invariant.");
2830 }
2831 }
2832 // Update birth stats for this block size.
2833 size_t num = fl->count();
2834 MutexLockerEx x(_indexedFreeListParLocks[word_sz],
2835 Mutex::_no_safepoint_check_flag);
2836 ssize_t births = _indexedFreeList[word_sz].split_births() + num;
2837 _indexedFreeList[word_sz].set_split_births(births);
2838 return;
2839 }
2840 }
2841 }
2842 // Otherwise, we'll split a block from the dictionary.
2843 FreeChunk* fc = NULL;
2844 FreeChunk* rem_fc = NULL;
2845 size_t rem;
2846 {
2847 MutexLockerEx x(parDictionaryAllocLock(),
2848 Mutex::_no_safepoint_check_flag);
2849 while (n > 0) {
2850 fc = dictionary()->get_chunk(MAX2(n * word_sz,
2851 _dictionary->min_size()),
2852 FreeBlockDictionary<FreeChunk>::atLeast);
2853 if (fc != NULL) {
2854 _bt.allocated((HeapWord*)fc, fc->size(), true /* reducing */); // update _unallocated_blk
2855 dictionary()->dict_census_udpate(fc->size(),
2856 true /*split*/,
2857 false /*birth*/);
2858 break;
2859 } else {
2860 n--;
2861 }
2862 }
2863 if (fc == NULL) return;
2864 // Otherwise, split up that block.
2865 assert((ssize_t)n >= 1, "Control point invariant");
2866 assert(fc->is_free(), "Error: should be a free block");
2867 _bt.verify_single_block((HeapWord*)fc, fc->size());
2868 const size_t nn = fc->size() / word_sz;
2869 n = MIN2(nn, n);
2870 assert((ssize_t)n >= 1, "Control point invariant");
2871 rem = fc->size() - n * word_sz;
2872 // If there is a remainder, and it's too small, allocate one fewer.
2873 if (rem > 0 && rem < MinChunkSize) {
2874 n--; rem += word_sz;
2875 }
2876 // Note that at this point we may have n == 0.
2877 assert((ssize_t)n >= 0, "Control point invariant");
2878
2879 // If n is 0, the chunk fc that was found is not large
2880 // enough to leave a viable remainder. We are unable to
2881 // allocate even one block. Return fc to the
2882 // dictionary and return, leaving "fl" empty.
2883 if (n == 0) {
2884 returnChunkToDictionary(fc);
2885 assert(fl->count() == 0, "We never allocated any blocks");
2886 return;
2887 }
2888
2889 // First return the remainder, if any.
2890 // Note that we hold the lock until we decide if we're going to give
2891 // back the remainder to the dictionary, since a concurrent allocation
2892 // may otherwise see the heap as empty. (We're willing to take that
2893 // hit if the block is a small block.)
2894 if (rem > 0) {
2895 size_t prefix_size = n * word_sz;
2896 rem_fc = (FreeChunk*)((HeapWord*)fc + prefix_size);
2897 rem_fc->set_size(rem);
2898 rem_fc->link_prev(NULL); // Mark as a free block for other (parallel) GC threads.
2899 rem_fc->link_next(NULL);
2900 // Above must occur before BOT is updated below.
2901 assert((ssize_t)n > 0 && prefix_size > 0 && rem_fc > fc, "Error");
2902 OrderAccess::storestore();
2903 _bt.split_block((HeapWord*)fc, fc->size(), prefix_size);
2904 assert(fc->is_free(), "Error");
2905 fc->set_size(prefix_size);
2906 if (rem >= IndexSetSize) {
2907 returnChunkToDictionary(rem_fc);
2908 dictionary()->dict_census_udpate(rem, true /*split*/, true /*birth*/);
2909 rem_fc = NULL;
2910 }
2911 // Otherwise, return it to the small list below.
2912 }
2913 }
2914 if (rem_fc != NULL) {
2915 MutexLockerEx x(_indexedFreeListParLocks[rem],
2916 Mutex::_no_safepoint_check_flag);
2917 _bt.verify_not_unallocated((HeapWord*)rem_fc, rem_fc->size());
2918 _indexedFreeList[rem].return_chunk_at_head(rem_fc);
2919 smallSplitBirth(rem);
2920 }
2921 assert((ssize_t)n > 0 && fc != NULL, "Consistency");
2922 // Now do the splitting up.
2923 // Must do this in reverse order, so that anybody attempting to
2924 // access the main chunk sees it as a single free block until we
2925 // change it.
2926 size_t fc_size = n * word_sz;
2927 // All but first chunk in this loop
2928 for (ssize_t i = n-1; i > 0; i--) {
2929 FreeChunk* ffc = (FreeChunk*)((HeapWord*)fc + i * word_sz);
2930 ffc->set_size(word_sz);
2931 ffc->link_prev(NULL); // Mark as a free block for other (parallel) GC threads.
2932 ffc->link_next(NULL);
2933 // Above must occur before BOT is updated below.
2934 OrderAccess::storestore();
2935 // splitting from the right, fc_size == (n - i + 1) * wordsize
2936 _bt.mark_block((HeapWord*)ffc, word_sz, true /* reducing */);
2937 fc_size -= word_sz;
2938 _bt.verify_not_unallocated((HeapWord*)ffc, ffc->size());
2939 _bt.verify_single_block((HeapWord*)ffc, ffc->size());
2940 _bt.verify_single_block((HeapWord*)fc, fc_size);
2941 // Push this on "fl".
2942 fl->return_chunk_at_head(ffc);
2943 }
2944 // First chunk
2945 assert(fc->is_free() && fc->size() == n*word_sz, "Error: should still be a free block");
2946 // The blocks above should show their new sizes before the first block below
2947 fc->set_size(word_sz);
2948 fc->link_prev(NULL); // idempotent wrt free-ness, see assert above
2949 fc->link_next(NULL);
2950 _bt.verify_not_unallocated((HeapWord*)fc, fc->size());
2951 _bt.verify_single_block((HeapWord*)fc, fc->size());
2952 fl->return_chunk_at_head(fc);
2953
2954 assert((ssize_t)n > 0 && (ssize_t)n == fl->count(), "Incorrect number of blocks");
2955 {
2956 // Update the stats for this block size.
2957 MutexLockerEx x(_indexedFreeListParLocks[word_sz],
2958 Mutex::_no_safepoint_check_flag);
2959 const ssize_t births = _indexedFreeList[word_sz].split_births() + n;
2960 _indexedFreeList[word_sz].set_split_births(births);
2961 // ssize_t new_surplus = _indexedFreeList[word_sz].surplus() + n;
2962 // _indexedFreeList[word_sz].set_surplus(new_surplus);
2963 }
2964
2965 // TRAP
2966 assert(fl->tail()->next() == NULL, "List invariant.");
2967 }
2968
2969 // Set up the space's par_seq_tasks structure for work claiming
2970 // for parallel rescan. See CMSParRemarkTask where this is currently used.
2971 // XXX Need to suitably abstract and generalize this and the next
2972 // method into one.
2973 void
2974 CompactibleFreeListSpace::
2975 initialize_sequential_subtasks_for_rescan(int n_threads) {
2976 // The "size" of each task is fixed according to rescan_task_size.
2977 assert(n_threads > 0, "Unexpected n_threads argument");
2978 const size_t task_size = rescan_task_size();
2979 size_t n_tasks = (used_region().word_size() + task_size - 1)/task_size;
2980 assert((n_tasks == 0) == used_region().is_empty(), "n_tasks incorrect");
2981 assert(n_tasks == 0 ||
2982 ((used_region().start() + (n_tasks - 1)*task_size < used_region().end()) &&
2983 (used_region().start() + n_tasks*task_size >= used_region().end())),
2984 "n_tasks calculation incorrect");
2985 SequentialSubTasksDone* pst = conc_par_seq_tasks();
2986 assert(!pst->valid(), "Clobbering existing data?");
2987 // Sets the condition for completion of the subtask (how many threads
2988 // need to finish in order to be done).
2989 pst->set_n_threads(n_threads);
2990 pst->set_n_tasks((int)n_tasks);
2991 }
2992
2993 // Set up the space's par_seq_tasks structure for work claiming
2994 // for parallel concurrent marking. See CMSConcMarkTask where this is currently used.
2995 void
2996 CompactibleFreeListSpace::
2997 initialize_sequential_subtasks_for_marking(int n_threads,
2998 HeapWord* low) {
2999 // The "size" of each task is fixed according to rescan_task_size.
3000 assert(n_threads > 0, "Unexpected n_threads argument");
3001 const size_t task_size = marking_task_size();
3002 assert(task_size > CardTableModRefBS::card_size_in_words &&
3003 (task_size % CardTableModRefBS::card_size_in_words == 0),
3004 "Otherwise arithmetic below would be incorrect");
3005 MemRegion span = _gen->reserved();
3006 if (low != NULL) {
3007 if (span.contains(low)) {
3008 // Align low down to a card boundary so that
3009 // we can use block_offset_careful() on span boundaries.
3010 HeapWord* aligned_low = (HeapWord*)align_size_down((uintptr_t)low,
3011 CardTableModRefBS::card_size);
3012 // Clip span prefix at aligned_low
3013 span = span.intersection(MemRegion(aligned_low, span.end()));
3014 } else if (low > span.end()) {
3015 span = MemRegion(low, low); // Null region
3016 } // else use entire span
3017 }
3018 assert(span.is_empty() ||
3019 ((uintptr_t)span.start() % CardTableModRefBS::card_size == 0),
3020 "span should start at a card boundary");
3021 size_t n_tasks = (span.word_size() + task_size - 1)/task_size;
3022 assert((n_tasks == 0) == span.is_empty(), "Inconsistency");
3023 assert(n_tasks == 0 ||
3024 ((span.start() + (n_tasks - 1)*task_size < span.end()) &&
3025 (span.start() + n_tasks*task_size >= span.end())),
3026 "n_tasks calculation incorrect");
3027 SequentialSubTasksDone* pst = conc_par_seq_tasks();
3028 assert(!pst->valid(), "Clobbering existing data?");
3029 // Sets the condition for completion of the subtask (how many threads
3030 // need to finish in order to be done).
3031 pst->set_n_threads(n_threads);
3032 pst->set_n_tasks((int)n_tasks);
3033 }