Mercurial > hg > truffle
annotate src/share/vm/gc_implementation/g1/concurrentMark.cpp @ 2951:0c0e407faa39
another fix to debug info (on-stack parameters), DCE removes unnecessary merges and LoopBegins whose LoopEnd went away
author | Lukas Stadler <lukas.stadler@jku.at> |
---|---|
date | Tue, 14 Jun 2011 16:47:31 +0200 |
parents | 8f1042ff784d |
children | cd8e33b2a8ad |
rev | line source |
---|---|
342 | 1 /* |
2149 | 2 * Copyright (c) 2001, 2011, Oracle and/or its affiliates. All rights reserved. |
342 | 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 * | |
1552
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
1547
diff
changeset
|
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
1547
diff
changeset
|
20 * or visit www.oracle.com if you need additional information or have any |
c18cbe5936b8
6941466: Oracle rebranding changes for Hotspot repositories
trims
parents:
1547
diff
changeset
|
21 * questions. |
342 | 22 * |
23 */ | |
24 | |
1972 | 25 #include "precompiled.hpp" |
26 #include "classfile/symbolTable.hpp" | |
27 #include "gc_implementation/g1/concurrentMark.hpp" | |
28 #include "gc_implementation/g1/concurrentMarkThread.inline.hpp" | |
29 #include "gc_implementation/g1/g1CollectedHeap.inline.hpp" | |
30 #include "gc_implementation/g1/g1CollectorPolicy.hpp" | |
31 #include "gc_implementation/g1/g1RemSet.hpp" | |
32 #include "gc_implementation/g1/heapRegionRemSet.hpp" | |
33 #include "gc_implementation/g1/heapRegionSeq.inline.hpp" | |
2125
7246a374a9f2
6458402: 3 jvmti tests fail with CMS and +ExplicitGCInvokesConcurrent
kamg
parents:
2037
diff
changeset
|
34 #include "gc_implementation/shared/vmGCOperations.hpp" |
1972 | 35 #include "memory/genOopClosures.inline.hpp" |
36 #include "memory/referencePolicy.hpp" | |
37 #include "memory/resourceArea.hpp" | |
38 #include "oops/oop.inline.hpp" | |
39 #include "runtime/handles.inline.hpp" | |
40 #include "runtime/java.hpp" | |
342 | 41 |
42 // | |
43 // CMS Bit Map Wrapper | |
44 | |
45 CMBitMapRO::CMBitMapRO(ReservedSpace rs, int shifter): | |
46 _bm((uintptr_t*)NULL,0), | |
47 _shifter(shifter) { | |
48 _bmStartWord = (HeapWord*)(rs.base()); | |
49 _bmWordSize = rs.size()/HeapWordSize; // rs.size() is in bytes | |
50 ReservedSpace brs(ReservedSpace::allocation_align_size_up( | |
51 (_bmWordSize >> (_shifter + LogBitsPerByte)) + 1)); | |
52 | |
53 guarantee(brs.is_reserved(), "couldn't allocate CMS bit map"); | |
54 // For now we'll just commit all of the bit map up fromt. | |
55 // Later on we'll try to be more parsimonious with swap. | |
56 guarantee(_virtual_space.initialize(brs, brs.size()), | |
57 "couldn't reseve backing store for CMS bit map"); | |
58 assert(_virtual_space.committed_size() == brs.size(), | |
59 "didn't reserve backing store for all of CMS bit map?"); | |
60 _bm.set_map((uintptr_t*)_virtual_space.low()); | |
61 assert(_virtual_space.committed_size() << (_shifter + LogBitsPerByte) >= | |
62 _bmWordSize, "inconsistency in bit map sizing"); | |
63 _bm.set_size(_bmWordSize >> _shifter); | |
64 } | |
65 | |
66 HeapWord* CMBitMapRO::getNextMarkedWordAddress(HeapWord* addr, | |
67 HeapWord* limit) const { | |
68 // First we must round addr *up* to a possible object boundary. | |
69 addr = (HeapWord*)align_size_up((intptr_t)addr, | |
70 HeapWordSize << _shifter); | |
71 size_t addrOffset = heapWordToOffset(addr); | |
72 if (limit == NULL) limit = _bmStartWord + _bmWordSize; | |
73 size_t limitOffset = heapWordToOffset(limit); | |
74 size_t nextOffset = _bm.get_next_one_offset(addrOffset, limitOffset); | |
75 HeapWord* nextAddr = offsetToHeapWord(nextOffset); | |
76 assert(nextAddr >= addr, "get_next_one postcondition"); | |
77 assert(nextAddr == limit || isMarked(nextAddr), | |
78 "get_next_one postcondition"); | |
79 return nextAddr; | |
80 } | |
81 | |
82 HeapWord* CMBitMapRO::getNextUnmarkedWordAddress(HeapWord* addr, | |
83 HeapWord* limit) const { | |
84 size_t addrOffset = heapWordToOffset(addr); | |
85 if (limit == NULL) limit = _bmStartWord + _bmWordSize; | |
86 size_t limitOffset = heapWordToOffset(limit); | |
87 size_t nextOffset = _bm.get_next_zero_offset(addrOffset, limitOffset); | |
88 HeapWord* nextAddr = offsetToHeapWord(nextOffset); | |
89 assert(nextAddr >= addr, "get_next_one postcondition"); | |
90 assert(nextAddr == limit || !isMarked(nextAddr), | |
91 "get_next_one postcondition"); | |
92 return nextAddr; | |
93 } | |
94 | |
95 int CMBitMapRO::heapWordDiffToOffsetDiff(size_t diff) const { | |
96 assert((diff & ((1 << _shifter) - 1)) == 0, "argument check"); | |
97 return (int) (diff >> _shifter); | |
98 } | |
99 | |
100 bool CMBitMapRO::iterate(BitMapClosure* cl, MemRegion mr) { | |
101 HeapWord* left = MAX2(_bmStartWord, mr.start()); | |
102 HeapWord* right = MIN2(_bmStartWord + _bmWordSize, mr.end()); | |
103 if (right > left) { | |
104 // Right-open interval [leftOffset, rightOffset). | |
105 return _bm.iterate(cl, heapWordToOffset(left), heapWordToOffset(right)); | |
106 } else { | |
107 return true; | |
108 } | |
109 } | |
110 | |
111 void CMBitMapRO::mostly_disjoint_range_union(BitMap* from_bitmap, | |
112 size_t from_start_index, | |
113 HeapWord* to_start_word, | |
114 size_t word_num) { | |
115 _bm.mostly_disjoint_range_union(from_bitmap, | |
116 from_start_index, | |
117 heapWordToOffset(to_start_word), | |
118 word_num); | |
119 } | |
120 | |
121 #ifndef PRODUCT | |
122 bool CMBitMapRO::covers(ReservedSpace rs) const { | |
123 // assert(_bm.map() == _virtual_space.low(), "map inconsistency"); | |
645
c3a720eefe82
6816308: Changes to allow builds with latest Windows SDK 6.1 on 64bit Windows 2003
kvn
parents:
619
diff
changeset
|
124 assert(((size_t)_bm.size() * (size_t)(1 << _shifter)) == _bmWordSize, |
342 | 125 "size inconsistency"); |
126 return _bmStartWord == (HeapWord*)(rs.base()) && | |
127 _bmWordSize == rs.size()>>LogHeapWordSize; | |
128 } | |
129 #endif | |
130 | |
131 void CMBitMap::clearAll() { | |
132 _bm.clear(); | |
133 return; | |
134 } | |
135 | |
136 void CMBitMap::markRange(MemRegion mr) { | |
137 mr.intersection(MemRegion(_bmStartWord, _bmWordSize)); | |
138 assert(!mr.is_empty(), "unexpected empty region"); | |
139 assert((offsetToHeapWord(heapWordToOffset(mr.end())) == | |
140 ((HeapWord *) mr.end())), | |
141 "markRange memory region end is not card aligned"); | |
142 // convert address range into offset range | |
143 _bm.at_put_range(heapWordToOffset(mr.start()), | |
144 heapWordToOffset(mr.end()), true); | |
145 } | |
146 | |
147 void CMBitMap::clearRange(MemRegion mr) { | |
148 mr.intersection(MemRegion(_bmStartWord, _bmWordSize)); | |
149 assert(!mr.is_empty(), "unexpected empty region"); | |
150 // convert address range into offset range | |
151 _bm.at_put_range(heapWordToOffset(mr.start()), | |
152 heapWordToOffset(mr.end()), false); | |
153 } | |
154 | |
155 MemRegion CMBitMap::getAndClearMarkedRegion(HeapWord* addr, | |
156 HeapWord* end_addr) { | |
157 HeapWord* start = getNextMarkedWordAddress(addr); | |
158 start = MIN2(start, end_addr); | |
159 HeapWord* end = getNextUnmarkedWordAddress(start); | |
160 end = MIN2(end, end_addr); | |
161 assert(start <= end, "Consistency check"); | |
162 MemRegion mr(start, end); | |
163 if (!mr.is_empty()) { | |
164 clearRange(mr); | |
165 } | |
166 return mr; | |
167 } | |
168 | |
169 CMMarkStack::CMMarkStack(ConcurrentMark* cm) : | |
170 _base(NULL), _cm(cm) | |
171 #ifdef ASSERT | |
172 , _drain_in_progress(false) | |
173 , _drain_in_progress_yields(false) | |
174 #endif | |
175 {} | |
176 | |
177 void CMMarkStack::allocate(size_t size) { | |
178 _base = NEW_C_HEAP_ARRAY(oop, size); | |
179 if (_base == NULL) | |
180 vm_exit_during_initialization("Failed to allocate " | |
181 "CM region mark stack"); | |
182 _index = 0; | |
183 // QQQQ cast ... | |
184 _capacity = (jint) size; | |
185 _oops_do_bound = -1; | |
186 NOT_PRODUCT(_max_depth = 0); | |
187 } | |
188 | |
189 CMMarkStack::~CMMarkStack() { | |
190 if (_base != NULL) FREE_C_HEAP_ARRAY(oop, _base); | |
191 } | |
192 | |
193 void CMMarkStack::par_push(oop ptr) { | |
194 while (true) { | |
195 if (isFull()) { | |
196 _overflow = true; | |
197 return; | |
198 } | |
199 // Otherwise... | |
200 jint index = _index; | |
201 jint next_index = index+1; | |
202 jint res = Atomic::cmpxchg(next_index, &_index, index); | |
203 if (res == index) { | |
204 _base[index] = ptr; | |
205 // Note that we don't maintain this atomically. We could, but it | |
206 // doesn't seem necessary. | |
207 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index)); | |
208 return; | |
209 } | |
210 // Otherwise, we need to try again. | |
211 } | |
212 } | |
213 | |
214 void CMMarkStack::par_adjoin_arr(oop* ptr_arr, int n) { | |
215 while (true) { | |
216 if (isFull()) { | |
217 _overflow = true; | |
218 return; | |
219 } | |
220 // Otherwise... | |
221 jint index = _index; | |
222 jint next_index = index + n; | |
223 if (next_index > _capacity) { | |
224 _overflow = true; | |
225 return; | |
226 } | |
227 jint res = Atomic::cmpxchg(next_index, &_index, index); | |
228 if (res == index) { | |
229 for (int i = 0; i < n; i++) { | |
230 int ind = index + i; | |
231 assert(ind < _capacity, "By overflow test above."); | |
232 _base[ind] = ptr_arr[i]; | |
233 } | |
234 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index)); | |
235 return; | |
236 } | |
237 // Otherwise, we need to try again. | |
238 } | |
239 } | |
240 | |
241 | |
242 void CMMarkStack::par_push_arr(oop* ptr_arr, int n) { | |
243 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); | |
244 jint start = _index; | |
245 jint next_index = start + n; | |
246 if (next_index > _capacity) { | |
247 _overflow = true; | |
248 return; | |
249 } | |
250 // Otherwise. | |
251 _index = next_index; | |
252 for (int i = 0; i < n; i++) { | |
253 int ind = start + i; | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
254 assert(ind < _capacity, "By overflow test above."); |
342 | 255 _base[ind] = ptr_arr[i]; |
256 } | |
257 } | |
258 | |
259 | |
260 bool CMMarkStack::par_pop_arr(oop* ptr_arr, int max, int* n) { | |
261 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); | |
262 jint index = _index; | |
263 if (index == 0) { | |
264 *n = 0; | |
265 return false; | |
266 } else { | |
267 int k = MIN2(max, index); | |
268 jint new_ind = index - k; | |
269 for (int j = 0; j < k; j++) { | |
270 ptr_arr[j] = _base[new_ind + j]; | |
271 } | |
272 _index = new_ind; | |
273 *n = k; | |
274 return true; | |
275 } | |
276 } | |
277 | |
278 | |
279 CMRegionStack::CMRegionStack() : _base(NULL) {} | |
280 | |
281 void CMRegionStack::allocate(size_t size) { | |
282 _base = NEW_C_HEAP_ARRAY(MemRegion, size); | |
283 if (_base == NULL) | |
284 vm_exit_during_initialization("Failed to allocate " | |
285 "CM region mark stack"); | |
286 _index = 0; | |
287 // QQQQ cast ... | |
288 _capacity = (jint) size; | |
289 } | |
290 | |
291 CMRegionStack::~CMRegionStack() { | |
292 if (_base != NULL) FREE_C_HEAP_ARRAY(oop, _base); | |
293 } | |
294 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
295 void CMRegionStack::push_lock_free(MemRegion mr) { |
342 | 296 assert(mr.word_size() > 0, "Precondition"); |
297 while (true) { | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
298 jint index = _index; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
299 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
300 if (index >= _capacity) { |
342 | 301 _overflow = true; |
302 return; | |
303 } | |
304 // Otherwise... | |
305 jint next_index = index+1; | |
306 jint res = Atomic::cmpxchg(next_index, &_index, index); | |
307 if (res == index) { | |
308 _base[index] = mr; | |
309 return; | |
310 } | |
311 // Otherwise, we need to try again. | |
312 } | |
313 } | |
314 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
315 // Lock-free pop of the region stack. Called during the concurrent |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
316 // marking / remark phases. Should only be called in tandem with |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
317 // other lock-free pops. |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
318 MemRegion CMRegionStack::pop_lock_free() { |
342 | 319 while (true) { |
320 jint index = _index; | |
321 | |
322 if (index == 0) { | |
323 return MemRegion(); | |
324 } | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
325 // Otherwise... |
342 | 326 jint next_index = index-1; |
327 jint res = Atomic::cmpxchg(next_index, &_index, index); | |
328 if (res == index) { | |
329 MemRegion mr = _base[next_index]; | |
330 if (mr.start() != NULL) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
331 assert(mr.end() != NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
332 assert(mr.word_size() > 0, "invariant"); |
342 | 333 return mr; |
334 } else { | |
335 // that entry was invalidated... let's skip it | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
336 assert(mr.end() == NULL, "invariant"); |
342 | 337 } |
338 } | |
339 // Otherwise, we need to try again. | |
340 } | |
341 } | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
342 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
343 #if 0 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
344 // The routines that manipulate the region stack with a lock are |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
345 // not currently used. They should be retained, however, as a |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
346 // diagnostic aid. |
1358
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
347 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
348 void CMRegionStack::push_with_lock(MemRegion mr) { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
349 assert(mr.word_size() > 0, "Precondition"); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
350 MutexLockerEx x(CMRegionStack_lock, Mutex::_no_safepoint_check_flag); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
351 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
352 if (isFull()) { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
353 _overflow = true; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
354 return; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
355 } |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
356 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
357 _base[_index] = mr; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
358 _index += 1; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
359 } |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
360 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
361 MemRegion CMRegionStack::pop_with_lock() { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
362 MutexLockerEx x(CMRegionStack_lock, Mutex::_no_safepoint_check_flag); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
363 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
364 while (true) { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
365 if (_index == 0) { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
366 return MemRegion(); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
367 } |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
368 _index -= 1; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
369 |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
370 MemRegion mr = _base[_index]; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
371 if (mr.start() != NULL) { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
372 assert(mr.end() != NULL, "invariant"); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
373 assert(mr.word_size() > 0, "invariant"); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
374 return mr; |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
375 } else { |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
376 // that entry was invalidated... let's skip it |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
377 assert(mr.end() == NULL, "invariant"); |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
378 } |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
379 } |
72f725c5a7be
6940310: G1: MT-unsafe calls to CM::region_stack_push() / CM::region_stack_pop()
tonyp
parents:
1317
diff
changeset
|
380 } |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
381 #endif |
342 | 382 |
383 bool CMRegionStack::invalidate_entries_into_cset() { | |
384 bool result = false; | |
385 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
386 for (int i = 0; i < _oops_do_bound; ++i) { | |
387 MemRegion mr = _base[i]; | |
388 if (mr.start() != NULL) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
389 assert(mr.end() != NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
390 assert(mr.word_size() > 0, "invariant"); |
342 | 391 HeapRegion* hr = g1h->heap_region_containing(mr.start()); |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
392 assert(hr != NULL, "invariant"); |
342 | 393 if (hr->in_collection_set()) { |
394 // The region points into the collection set | |
395 _base[i] = MemRegion(); | |
396 result = true; | |
397 } | |
398 } else { | |
399 // that entry was invalidated... let's skip it | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
400 assert(mr.end() == NULL, "invariant"); |
342 | 401 } |
402 } | |
403 return result; | |
404 } | |
405 | |
406 template<class OopClosureClass> | |
407 bool CMMarkStack::drain(OopClosureClass* cl, CMBitMap* bm, bool yield_after) { | |
408 assert(!_drain_in_progress || !_drain_in_progress_yields || yield_after | |
409 || SafepointSynchronize::is_at_safepoint(), | |
410 "Drain recursion must be yield-safe."); | |
411 bool res = true; | |
412 debug_only(_drain_in_progress = true); | |
413 debug_only(_drain_in_progress_yields = yield_after); | |
414 while (!isEmpty()) { | |
415 oop newOop = pop(); | |
416 assert(G1CollectedHeap::heap()->is_in_reserved(newOop), "Bad pop"); | |
417 assert(newOop->is_oop(), "Expected an oop"); | |
418 assert(bm == NULL || bm->isMarked((HeapWord*)newOop), | |
419 "only grey objects on this stack"); | |
420 // iterate over the oops in this oop, marking and pushing | |
421 // the ones in CMS generation. | |
422 newOop->oop_iterate(cl); | |
423 if (yield_after && _cm->do_yield_check()) { | |
424 res = false; break; | |
425 } | |
426 } | |
427 debug_only(_drain_in_progress = false); | |
428 return res; | |
429 } | |
430 | |
431 void CMMarkStack::oops_do(OopClosure* f) { | |
432 if (_index == 0) return; | |
433 assert(_oops_do_bound != -1 && _oops_do_bound <= _index, | |
434 "Bound must be set."); | |
435 for (int i = 0; i < _oops_do_bound; i++) { | |
436 f->do_oop(&_base[i]); | |
437 } | |
438 _oops_do_bound = -1; | |
439 } | |
440 | |
441 bool ConcurrentMark::not_yet_marked(oop obj) const { | |
442 return (_g1h->is_obj_ill(obj) | |
443 || (_g1h->is_in_permanent(obj) | |
444 && !nextMarkBitMap()->isMarked((HeapWord*)obj))); | |
445 } | |
446 | |
447 #ifdef _MSC_VER // the use of 'this' below gets a warning, make it go away | |
448 #pragma warning( disable:4355 ) // 'this' : used in base member initializer list | |
449 #endif // _MSC_VER | |
450 | |
451 ConcurrentMark::ConcurrentMark(ReservedSpace rs, | |
452 int max_regions) : | |
453 _markBitMap1(rs, MinObjAlignment - 1), | |
454 _markBitMap2(rs, MinObjAlignment - 1), | |
455 | |
456 _parallel_marking_threads(0), | |
457 _sleep_factor(0.0), | |
458 _marking_task_overhead(1.0), | |
459 _cleanup_sleep_factor(0.0), | |
460 _cleanup_task_overhead(1.0), | |
2152 | 461 _cleanup_list("Cleanup List"), |
342 | 462 _region_bm(max_regions, false /* in_resource_area*/), |
463 _card_bm((rs.size() + CardTableModRefBS::card_size - 1) >> | |
464 CardTableModRefBS::card_shift, | |
465 false /* in_resource_area*/), | |
466 _prevMarkBitMap(&_markBitMap1), | |
467 _nextMarkBitMap(&_markBitMap2), | |
468 _at_least_one_mark_complete(false), | |
469 | |
470 _markStack(this), | |
471 _regionStack(), | |
472 // _finger set in set_non_marking_state | |
473 | |
474 _max_task_num(MAX2(ParallelGCThreads, (size_t)1)), | |
475 // _active_tasks set in set_non_marking_state | |
476 // _tasks set inside the constructor | |
477 _task_queues(new CMTaskQueueSet((int) _max_task_num)), | |
478 _terminator(ParallelTaskTerminator((int) _max_task_num, _task_queues)), | |
479 | |
480 _has_overflown(false), | |
481 _concurrent(false), | |
619
7ea5ca260b28
6814467: G1: small fixes related to concurrent marking verboseness
tonyp
parents:
470
diff
changeset
|
482 _has_aborted(false), |
7ea5ca260b28
6814467: G1: small fixes related to concurrent marking verboseness
tonyp
parents:
470
diff
changeset
|
483 _restart_for_overflow(false), |
7ea5ca260b28
6814467: G1: small fixes related to concurrent marking verboseness
tonyp
parents:
470
diff
changeset
|
484 _concurrent_marking_in_progress(false), |
7ea5ca260b28
6814467: G1: small fixes related to concurrent marking verboseness
tonyp
parents:
470
diff
changeset
|
485 _should_gray_objects(false), |
342 | 486 |
487 // _verbose_level set below | |
488 | |
489 _init_times(), | |
490 _remark_times(), _remark_mark_times(), _remark_weak_ref_times(), | |
491 _cleanup_times(), | |
492 _total_counting_time(0.0), | |
493 _total_rs_scrub_time(0.0), | |
494 | |
936 | 495 _parallel_workers(NULL) |
342 | 496 { |
497 CMVerboseLevel verbose_level = | |
498 (CMVerboseLevel) G1MarkingVerboseLevel; | |
499 if (verbose_level < no_verbose) | |
500 verbose_level = no_verbose; | |
501 if (verbose_level > high_verbose) | |
502 verbose_level = high_verbose; | |
503 _verbose_level = verbose_level; | |
504 | |
505 if (verbose_low()) | |
506 gclog_or_tty->print_cr("[global] init, heap start = "PTR_FORMAT", " | |
507 "heap end = "PTR_FORMAT, _heap_start, _heap_end); | |
508 | |
1284 | 509 _markStack.allocate(MarkStackSize); |
751 | 510 _regionStack.allocate(G1MarkRegionStackSize); |
342 | 511 |
512 // Create & start a ConcurrentMark thread. | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
513 _cmThread = new ConcurrentMarkThread(this); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
514 assert(cmThread() != NULL, "CM Thread should have been created"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
515 assert(cmThread()->cm() != NULL, "CM Thread should refer to this cm"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
516 |
342 | 517 _g1h = G1CollectedHeap::heap(); |
518 assert(CGC_lock != NULL, "Where's the CGC_lock?"); | |
519 assert(_markBitMap1.covers(rs), "_markBitMap1 inconsistency"); | |
520 assert(_markBitMap2.covers(rs), "_markBitMap2 inconsistency"); | |
521 | |
522 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set(); | |
1282 | 523 satb_qs.set_buffer_size(G1SATBBufferSize); |
342 | 524 |
525 _tasks = NEW_C_HEAP_ARRAY(CMTask*, _max_task_num); | |
526 _accum_task_vtime = NEW_C_HEAP_ARRAY(double, _max_task_num); | |
527 | |
528 // so that the assertion in MarkingTaskQueue::task_queue doesn't fail | |
529 _active_tasks = _max_task_num; | |
530 for (int i = 0; i < (int) _max_task_num; ++i) { | |
531 CMTaskQueue* task_queue = new CMTaskQueue(); | |
532 task_queue->initialize(); | |
533 _task_queues->register_queue(i, task_queue); | |
534 | |
535 _tasks[i] = new CMTask(i, this, task_queue, _task_queues); | |
536 _accum_task_vtime[i] = 0.0; | |
537 } | |
538 | |
1284 | 539 if (ConcGCThreads > ParallelGCThreads) { |
540 vm_exit_during_initialization("Can't have more ConcGCThreads " | |
342 | 541 "than ParallelGCThreads."); |
542 } | |
543 if (ParallelGCThreads == 0) { | |
544 // if we are not running with any parallel GC threads we will not | |
545 // spawn any marking threads either | |
546 _parallel_marking_threads = 0; | |
547 _sleep_factor = 0.0; | |
548 _marking_task_overhead = 1.0; | |
549 } else { | |
1284 | 550 if (ConcGCThreads > 0) { |
551 // notice that ConcGCThreads overwrites G1MarkingOverheadPercent | |
342 | 552 // if both are set |
553 | |
1284 | 554 _parallel_marking_threads = ConcGCThreads; |
342 | 555 _sleep_factor = 0.0; |
556 _marking_task_overhead = 1.0; | |
751 | 557 } else if (G1MarkingOverheadPercent > 0) { |
342 | 558 // we will calculate the number of parallel marking threads |
559 // based on a target overhead with respect to the soft real-time | |
560 // goal | |
561 | |
751 | 562 double marking_overhead = (double) G1MarkingOverheadPercent / 100.0; |
342 | 563 double overall_cm_overhead = |
751 | 564 (double) MaxGCPauseMillis * marking_overhead / |
565 (double) GCPauseIntervalMillis; | |
342 | 566 double cpu_ratio = 1.0 / (double) os::processor_count(); |
567 double marking_thread_num = ceil(overall_cm_overhead / cpu_ratio); | |
568 double marking_task_overhead = | |
569 overall_cm_overhead / marking_thread_num * | |
570 (double) os::processor_count(); | |
571 double sleep_factor = | |
572 (1.0 - marking_task_overhead) / marking_task_overhead; | |
573 | |
574 _parallel_marking_threads = (size_t) marking_thread_num; | |
575 _sleep_factor = sleep_factor; | |
576 _marking_task_overhead = marking_task_overhead; | |
577 } else { | |
578 _parallel_marking_threads = MAX2((ParallelGCThreads + 2) / 4, (size_t)1); | |
579 _sleep_factor = 0.0; | |
580 _marking_task_overhead = 1.0; | |
581 } | |
582 | |
583 if (parallel_marking_threads() > 1) | |
584 _cleanup_task_overhead = 1.0; | |
585 else | |
586 _cleanup_task_overhead = marking_task_overhead(); | |
587 _cleanup_sleep_factor = | |
588 (1.0 - cleanup_task_overhead()) / cleanup_task_overhead(); | |
589 | |
590 #if 0 | |
591 gclog_or_tty->print_cr("Marking Threads %d", parallel_marking_threads()); | |
592 gclog_or_tty->print_cr("CM Marking Task Overhead %1.4lf", marking_task_overhead()); | |
593 gclog_or_tty->print_cr("CM Sleep Factor %1.4lf", sleep_factor()); | |
594 gclog_or_tty->print_cr("CL Marking Task Overhead %1.4lf", cleanup_task_overhead()); | |
595 gclog_or_tty->print_cr("CL Sleep Factor %1.4lf", cleanup_sleep_factor()); | |
596 #endif | |
597 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
598 guarantee(parallel_marking_threads() > 0, "peace of mind"); |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
599 _parallel_workers = new FlexibleWorkGang("G1 Parallel Marking Threads", |
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
600 (int) _parallel_marking_threads, false, true); |
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
601 if (_parallel_workers == NULL) { |
342 | 602 vm_exit_during_initialization("Failed necessary allocation."); |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
603 } else { |
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
604 _parallel_workers->initialize_workers(); |
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
605 } |
342 | 606 } |
607 | |
608 // so that the call below can read a sensible value | |
609 _heap_start = (HeapWord*) rs.base(); | |
610 set_non_marking_state(); | |
611 } | |
612 | |
613 void ConcurrentMark::update_g1_committed(bool force) { | |
614 // If concurrent marking is not in progress, then we do not need to | |
615 // update _heap_end. This has a subtle and important | |
616 // side-effect. Imagine that two evacuation pauses happen between | |
617 // marking completion and remark. The first one can grow the | |
618 // heap (hence now the finger is below the heap end). Then, the | |
619 // second one could unnecessarily push regions on the region | |
620 // stack. This causes the invariant that the region stack is empty | |
621 // at the beginning of remark to be false. By ensuring that we do | |
622 // not observe heap expansions after marking is complete, then we do | |
623 // not have this problem. | |
624 if (!concurrent_marking_in_progress() && !force) | |
625 return; | |
626 | |
627 MemRegion committed = _g1h->g1_committed(); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
628 assert(committed.start() == _heap_start, "start shouldn't change"); |
342 | 629 HeapWord* new_end = committed.end(); |
630 if (new_end > _heap_end) { | |
631 // The heap has been expanded. | |
632 | |
633 _heap_end = new_end; | |
634 } | |
635 // Notice that the heap can also shrink. However, this only happens | |
636 // during a Full GC (at least currently) and the entire marking | |
637 // phase will bail out and the task will not be restarted. So, let's | |
638 // do nothing. | |
639 } | |
640 | |
641 void ConcurrentMark::reset() { | |
642 // Starting values for these two. This should be called in a STW | |
643 // phase. CM will be notified of any future g1_committed expansions | |
644 // will be at the end of evacuation pauses, when tasks are | |
645 // inactive. | |
646 MemRegion committed = _g1h->g1_committed(); | |
647 _heap_start = committed.start(); | |
648 _heap_end = committed.end(); | |
649 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
650 // Separated the asserts so that we know which one fires. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
651 assert(_heap_start != NULL, "heap bounds should look ok"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
652 assert(_heap_end != NULL, "heap bounds should look ok"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
653 assert(_heap_start < _heap_end, "heap bounds should look ok"); |
342 | 654 |
655 // reset all the marking data structures and any necessary flags | |
656 clear_marking_state(); | |
657 | |
658 if (verbose_low()) | |
659 gclog_or_tty->print_cr("[global] resetting"); | |
660 | |
661 // We do reset all of them, since different phases will use | |
662 // different number of active threads. So, it's easiest to have all | |
663 // of them ready. | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
664 for (int i = 0; i < (int) _max_task_num; ++i) { |
342 | 665 _tasks[i]->reset(_nextMarkBitMap); |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
666 } |
342 | 667 |
668 // we need this to make sure that the flag is on during the evac | |
669 // pause with initial mark piggy-backed | |
670 set_concurrent_marking_in_progress(); | |
671 } | |
672 | |
673 void ConcurrentMark::set_phase(size_t active_tasks, bool concurrent) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
674 assert(active_tasks <= _max_task_num, "we should not have more"); |
342 | 675 |
676 _active_tasks = active_tasks; | |
677 // Need to update the three data structures below according to the | |
678 // number of active threads for this phase. | |
679 _terminator = ParallelTaskTerminator((int) active_tasks, _task_queues); | |
680 _first_overflow_barrier_sync.set_n_workers((int) active_tasks); | |
681 _second_overflow_barrier_sync.set_n_workers((int) active_tasks); | |
682 | |
683 _concurrent = concurrent; | |
684 // We propagate this to all tasks, not just the active ones. | |
685 for (int i = 0; i < (int) _max_task_num; ++i) | |
686 _tasks[i]->set_concurrent(concurrent); | |
687 | |
688 if (concurrent) { | |
689 set_concurrent_marking_in_progress(); | |
690 } else { | |
691 // We currently assume that the concurrent flag has been set to | |
692 // false before we start remark. At this point we should also be | |
693 // in a STW phase. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
694 assert(!concurrent_marking_in_progress(), "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
695 assert(_finger == _heap_end, "only way to get here"); |
342 | 696 update_g1_committed(true); |
697 } | |
698 } | |
699 | |
700 void ConcurrentMark::set_non_marking_state() { | |
701 // We set the global marking state to some default values when we're | |
702 // not doing marking. | |
703 clear_marking_state(); | |
704 _active_tasks = 0; | |
705 clear_concurrent_marking_in_progress(); | |
706 } | |
707 | |
708 ConcurrentMark::~ConcurrentMark() { | |
709 for (int i = 0; i < (int) _max_task_num; ++i) { | |
710 delete _task_queues->queue(i); | |
711 delete _tasks[i]; | |
712 } | |
713 delete _task_queues; | |
714 FREE_C_HEAP_ARRAY(CMTask*, _max_task_num); | |
715 } | |
716 | |
717 // This closure is used to mark refs into the g1 generation | |
718 // from external roots in the CMS bit map. | |
719 // Called at the first checkpoint. | |
720 // | |
721 | |
722 void ConcurrentMark::clearNextBitmap() { | |
1359
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
723 G1CollectedHeap* g1h = G1CollectedHeap::heap(); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
724 G1CollectorPolicy* g1p = g1h->g1_policy(); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
725 |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
726 // Make sure that the concurrent mark thread looks to still be in |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
727 // the current cycle. |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
728 guarantee(cmThread()->during_cycle(), "invariant"); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
729 |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
730 // We are finishing up the current cycle by clearing the next |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
731 // marking bitmap and getting it ready for the next cycle. During |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
732 // this time no other cycle can start. So, let's make sure that this |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
733 // is the case. |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
734 guarantee(!g1h->mark_in_progress(), "invariant"); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
735 |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
736 // clear the mark bitmap (no grey objects to start with). |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
737 // We need to do this in chunks and offer to yield in between |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
738 // each chunk. |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
739 HeapWord* start = _nextMarkBitMap->startWord(); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
740 HeapWord* end = _nextMarkBitMap->endWord(); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
741 HeapWord* cur = start; |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
742 size_t chunkSize = M; |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
743 while (cur < end) { |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
744 HeapWord* next = cur + chunkSize; |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
745 if (next > end) |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
746 next = end; |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
747 MemRegion mr(cur,next); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
748 _nextMarkBitMap->clearRange(mr); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
749 cur = next; |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
750 do_yield_check(); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
751 |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
752 // Repeat the asserts from above. We'll do them as asserts here to |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
753 // minimize their overhead on the product. However, we'll have |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
754 // them as guarantees at the beginning / end of the bitmap |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
755 // clearing to get some checking in the product. |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
756 assert(cmThread()->during_cycle(), "invariant"); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
757 assert(!g1h->mark_in_progress(), "invariant"); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
758 } |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
759 |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
760 // Repeat the asserts from above. |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
761 guarantee(cmThread()->during_cycle(), "invariant"); |
23b1b27ac76c
6909756: G1: guarantee(G1CollectedHeap::heap()->mark_in_progress(),"Precondition.")
tonyp
parents:
1358
diff
changeset
|
762 guarantee(!g1h->mark_in_progress(), "invariant"); |
342 | 763 } |
764 | |
765 class NoteStartOfMarkHRClosure: public HeapRegionClosure { | |
766 public: | |
767 bool doHeapRegion(HeapRegion* r) { | |
768 if (!r->continuesHumongous()) { | |
769 r->note_start_of_marking(true); | |
770 } | |
771 return false; | |
772 } | |
773 }; | |
774 | |
775 void ConcurrentMark::checkpointRootsInitialPre() { | |
776 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
777 G1CollectorPolicy* g1p = g1h->g1_policy(); | |
778 | |
779 _has_aborted = false; | |
780 | |
1547
fb1a39993f69
6951319: enable solaris builds using Sun Studio 12 update 1
jcoomes
parents:
1388
diff
changeset
|
781 #ifndef PRODUCT |
1044 | 782 if (G1PrintReachableAtInitialMark) { |
1388 | 783 print_reachable("at-cycle-start", |
784 true /* use_prev_marking */, true /* all */); | |
1044 | 785 } |
1547
fb1a39993f69
6951319: enable solaris builds using Sun Studio 12 update 1
jcoomes
parents:
1388
diff
changeset
|
786 #endif |
342 | 787 |
788 // Initialise marking structures. This has to be done in a STW phase. | |
789 reset(); | |
790 } | |
791 | |
792 class CMMarkRootsClosure: public OopsInGenClosure { | |
793 private: | |
794 ConcurrentMark* _cm; | |
795 G1CollectedHeap* _g1h; | |
796 bool _do_barrier; | |
797 | |
798 public: | |
799 CMMarkRootsClosure(ConcurrentMark* cm, | |
800 G1CollectedHeap* g1h, | |
801 bool do_barrier) : _cm(cm), _g1h(g1h), | |
802 _do_barrier(do_barrier) { } | |
803 | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
804 virtual void do_oop(narrowOop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
805 virtual void do_oop( oop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
806 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
807 template <class T> void do_oop_work(T* p) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
808 T heap_oop = oopDesc::load_heap_oop(p); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
809 if (!oopDesc::is_null(heap_oop)) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
810 oop obj = oopDesc::decode_heap_oop_not_null(heap_oop); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
811 assert(obj->is_oop() || obj->mark() == NULL, |
342 | 812 "expected an oop, possibly with mark word displaced"); |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
813 HeapWord* addr = (HeapWord*)obj; |
342 | 814 if (_g1h->is_in_g1_reserved(addr)) { |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
815 _cm->grayRoot(obj); |
342 | 816 } |
817 } | |
818 if (_do_barrier) { | |
819 assert(!_g1h->is_in_g1_reserved(p), | |
820 "Should be called on external roots"); | |
821 do_barrier(p); | |
822 } | |
823 } | |
824 }; | |
825 | |
826 void ConcurrentMark::checkpointRootsInitialPost() { | |
827 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
828 | |
829 // For each region note start of marking. | |
830 NoteStartOfMarkHRClosure startcl; | |
831 g1h->heap_region_iterate(&startcl); | |
832 | |
833 // Start weak-reference discovery. | |
834 ReferenceProcessor* rp = g1h->ref_processor(); | |
835 rp->verify_no_references_recorded(); | |
836 rp->enable_discovery(); // enable ("weak") refs discovery | |
457
27a80744a83b
6778647: snap(), snap_policy() should be renamed setup(), setup_policy()
ysr
parents:
453
diff
changeset
|
837 rp->setup_policy(false); // snapshot the soft ref policy to be used in this cycle |
342 | 838 |
839 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); | |
1317
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
840 // This is the start of the marking cycle, we're expected all |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
841 // threads to have SATB queues with active set to false. |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
842 satb_mq_set.set_active_all_threads(true, /* new active value */ |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
843 false /* expected_active */); |
342 | 844 |
845 // update_g1_committed() will be called at the end of an evac pause | |
846 // when marking is on. So, it's also called at the end of the | |
847 // initial-mark pause to update the heap end, if the heap expands | |
848 // during it. No need to call it here. | |
849 } | |
850 | |
851 // Checkpoint the roots into this generation from outside | |
852 // this generation. [Note this initial checkpoint need only | |
853 // be approximate -- we'll do a catch up phase subsequently.] | |
854 void ConcurrentMark::checkpointRootsInitial() { | |
855 assert(SafepointSynchronize::is_at_safepoint(), "world should be stopped"); | |
856 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
857 | |
858 double start = os::elapsedTime(); | |
859 | |
860 G1CollectorPolicy* g1p = G1CollectedHeap::heap()->g1_policy(); | |
861 g1p->record_concurrent_mark_init_start(); | |
862 checkpointRootsInitialPre(); | |
863 | |
864 // YSR: when concurrent precleaning is in place, we'll | |
865 // need to clear the cached card table here | |
866 | |
867 ResourceMark rm; | |
868 HandleMark hm; | |
869 | |
870 g1h->ensure_parsability(false); | |
871 g1h->perm_gen()->save_marks(); | |
872 | |
873 CMMarkRootsClosure notOlder(this, g1h, false); | |
874 CMMarkRootsClosure older(this, g1h, true); | |
875 | |
876 g1h->set_marking_started(); | |
877 g1h->rem_set()->prepare_for_younger_refs_iterate(false); | |
878 | |
989
148e5441d916
6863023: need non-perm oops in code cache for JSR 292
jrose
parents:
912
diff
changeset
|
879 g1h->process_strong_roots(true, // activate StrongRootsScope |
148e5441d916
6863023: need non-perm oops in code cache for JSR 292
jrose
parents:
912
diff
changeset
|
880 false, // fake perm gen collection |
342 | 881 SharedHeap::SO_AllClasses, |
882 ¬Older, // Regular roots | |
989
148e5441d916
6863023: need non-perm oops in code cache for JSR 292
jrose
parents:
912
diff
changeset
|
883 NULL, // do not visit active blobs |
342 | 884 &older // Perm Gen Roots |
885 ); | |
886 checkpointRootsInitialPost(); | |
887 | |
888 // Statistics. | |
889 double end = os::elapsedTime(); | |
890 _init_times.add((end - start) * 1000.0); | |
891 | |
892 g1p->record_concurrent_mark_init_end(); | |
893 } | |
894 | |
895 /* | |
896 Notice that in the next two methods, we actually leave the STS | |
897 during the barrier sync and join it immediately afterwards. If we | |
898 do not do this, this then the following deadlock can occur: one | |
899 thread could be in the barrier sync code, waiting for the other | |
900 thread to also sync up, whereas another one could be trying to | |
901 yield, while also waiting for the other threads to sync up too. | |
902 | |
903 Because the thread that does the sync barrier has left the STS, it | |
904 is possible to be suspended for a Full GC or an evacuation pause | |
905 could occur. This is actually safe, since the entering the sync | |
906 barrier is one of the last things do_marking_step() does, and it | |
907 doesn't manipulate any data structures afterwards. | |
908 */ | |
909 | |
910 void ConcurrentMark::enter_first_sync_barrier(int task_num) { | |
911 if (verbose_low()) | |
912 gclog_or_tty->print_cr("[%d] entering first barrier", task_num); | |
913 | |
914 ConcurrentGCThread::stsLeave(); | |
915 _first_overflow_barrier_sync.enter(); | |
916 ConcurrentGCThread::stsJoin(); | |
917 // at this point everyone should have synced up and not be doing any | |
918 // more work | |
919 | |
920 if (verbose_low()) | |
921 gclog_or_tty->print_cr("[%d] leaving first barrier", task_num); | |
922 | |
923 // let task 0 do this | |
924 if (task_num == 0) { | |
925 // task 0 is responsible for clearing the global data structures | |
926 clear_marking_state(); | |
927 | |
928 if (PrintGC) { | |
929 gclog_or_tty->date_stamp(PrintGCDateStamps); | |
930 gclog_or_tty->stamp(PrintGCTimeStamps); | |
931 gclog_or_tty->print_cr("[GC concurrent-mark-reset-for-overflow]"); | |
932 } | |
933 } | |
934 | |
935 // after this, each task should reset its own data structures then | |
936 // then go into the second barrier | |
937 } | |
938 | |
939 void ConcurrentMark::enter_second_sync_barrier(int task_num) { | |
940 if (verbose_low()) | |
941 gclog_or_tty->print_cr("[%d] entering second barrier", task_num); | |
942 | |
943 ConcurrentGCThread::stsLeave(); | |
944 _second_overflow_barrier_sync.enter(); | |
945 ConcurrentGCThread::stsJoin(); | |
946 // at this point everything should be re-initialised and ready to go | |
947 | |
948 if (verbose_low()) | |
949 gclog_or_tty->print_cr("[%d] leaving second barrier", task_num); | |
950 } | |
951 | |
952 void ConcurrentMark::grayRoot(oop p) { | |
953 HeapWord* addr = (HeapWord*) p; | |
954 // We can't really check against _heap_start and _heap_end, since it | |
955 // is possible during an evacuation pause with piggy-backed | |
956 // initial-mark that the committed space is expanded during the | |
957 // pause without CM observing this change. So the assertions below | |
958 // is a bit conservative; but better than nothing. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
959 assert(_g1h->g1_committed().contains(addr), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
960 "address should be within the heap bounds"); |
342 | 961 |
962 if (!_nextMarkBitMap->isMarked(addr)) | |
963 _nextMarkBitMap->parMark(addr); | |
964 } | |
965 | |
966 void ConcurrentMark::grayRegionIfNecessary(MemRegion mr) { | |
967 // The objects on the region have already been marked "in bulk" by | |
968 // the caller. We only need to decide whether to push the region on | |
969 // the region stack or not. | |
970 | |
971 if (!concurrent_marking_in_progress() || !_should_gray_objects) | |
972 // We're done with marking and waiting for remark. We do not need to | |
973 // push anything else on the region stack. | |
974 return; | |
975 | |
976 HeapWord* finger = _finger; | |
977 | |
978 if (verbose_low()) | |
979 gclog_or_tty->print_cr("[global] attempting to push " | |
980 "region ["PTR_FORMAT", "PTR_FORMAT"), finger is at " | |
981 PTR_FORMAT, mr.start(), mr.end(), finger); | |
982 | |
983 if (mr.start() < finger) { | |
984 // The finger is always heap region aligned and it is not possible | |
985 // for mr to span heap regions. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
986 assert(mr.end() <= finger, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
987 |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
988 // Separated the asserts so that we know which one fires. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
989 assert(mr.start() <= mr.end(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
990 "region boundaries should fall within the committed space"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
991 assert(_heap_start <= mr.start(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
992 "region boundaries should fall within the committed space"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
993 assert(mr.end() <= _heap_end, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
994 "region boundaries should fall within the committed space"); |
342 | 995 if (verbose_low()) |
996 gclog_or_tty->print_cr("[global] region ["PTR_FORMAT", "PTR_FORMAT") " | |
997 "below the finger, pushing it", | |
998 mr.start(), mr.end()); | |
999 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
1000 if (!region_stack_push_lock_free(mr)) { |
342 | 1001 if (verbose_low()) |
1002 gclog_or_tty->print_cr("[global] region stack has overflown."); | |
1003 } | |
1004 } | |
1005 } | |
1006 | |
1007 void ConcurrentMark::markAndGrayObjectIfNecessary(oop p) { | |
1008 // The object is not marked by the caller. We need to at least mark | |
1009 // it and maybe push in on the stack. | |
1010 | |
1011 HeapWord* addr = (HeapWord*)p; | |
1012 if (!_nextMarkBitMap->isMarked(addr)) { | |
1013 // We definitely need to mark it, irrespective whether we bail out | |
1014 // because we're done with marking. | |
1015 if (_nextMarkBitMap->parMark(addr)) { | |
1016 if (!concurrent_marking_in_progress() || !_should_gray_objects) | |
1017 // If we're done with concurrent marking and we're waiting for | |
1018 // remark, then we're not pushing anything on the stack. | |
1019 return; | |
1020 | |
1021 // No OrderAccess:store_load() is needed. It is implicit in the | |
1022 // CAS done in parMark(addr) above | |
1023 HeapWord* finger = _finger; | |
1024 | |
1025 if (addr < finger) { | |
1026 if (!mark_stack_push(oop(addr))) { | |
1027 if (verbose_low()) | |
1028 gclog_or_tty->print_cr("[global] global stack overflow " | |
1029 "during parMark"); | |
1030 } | |
1031 } | |
1032 } | |
1033 } | |
1034 } | |
1035 | |
1036 class CMConcurrentMarkingTask: public AbstractGangTask { | |
1037 private: | |
1038 ConcurrentMark* _cm; | |
1039 ConcurrentMarkThread* _cmt; | |
1040 | |
1041 public: | |
1042 void work(int worker_i) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1043 assert(Thread::current()->is_ConcurrentGC_thread(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1044 "this should only be done by a conc GC thread"); |
1974
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
1045 ResourceMark rm; |
342 | 1046 |
1047 double start_vtime = os::elapsedVTime(); | |
1048 | |
1049 ConcurrentGCThread::stsJoin(); | |
1050 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1051 assert((size_t) worker_i < _cm->active_tasks(), "invariant"); |
342 | 1052 CMTask* the_task = _cm->task(worker_i); |
1053 the_task->record_start_time(); | |
1054 if (!_cm->has_aborted()) { | |
1055 do { | |
1056 double start_vtime_sec = os::elapsedVTime(); | |
1057 double start_time_sec = os::elapsedTime(); | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1058 double mark_step_duration_ms = G1ConcMarkStepDurationMillis; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1059 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1060 the_task->do_marking_step(mark_step_duration_ms, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1061 true /* do_stealing */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1062 true /* do_termination */); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1063 |
342 | 1064 double end_time_sec = os::elapsedTime(); |
1065 double end_vtime_sec = os::elapsedVTime(); | |
1066 double elapsed_vtime_sec = end_vtime_sec - start_vtime_sec; | |
1067 double elapsed_time_sec = end_time_sec - start_time_sec; | |
1068 _cm->clear_has_overflown(); | |
1069 | |
1070 bool ret = _cm->do_yield_check(worker_i); | |
1071 | |
1072 jlong sleep_time_ms; | |
1073 if (!_cm->has_aborted() && the_task->has_aborted()) { | |
1074 sleep_time_ms = | |
1075 (jlong) (elapsed_vtime_sec * _cm->sleep_factor() * 1000.0); | |
1076 ConcurrentGCThread::stsLeave(); | |
1077 os::sleep(Thread::current(), sleep_time_ms, false); | |
1078 ConcurrentGCThread::stsJoin(); | |
1079 } | |
1080 double end_time2_sec = os::elapsedTime(); | |
1081 double elapsed_time2_sec = end_time2_sec - start_time_sec; | |
1082 | |
1083 #if 0 | |
1084 gclog_or_tty->print_cr("CM: elapsed %1.4lf ms, sleep %1.4lf ms, " | |
1085 "overhead %1.4lf", | |
1086 elapsed_vtime_sec * 1000.0, (double) sleep_time_ms, | |
1087 the_task->conc_overhead(os::elapsedTime()) * 8.0); | |
1088 gclog_or_tty->print_cr("elapsed time %1.4lf ms, time 2: %1.4lf ms", | |
1089 elapsed_time_sec * 1000.0, elapsed_time2_sec * 1000.0); | |
1090 #endif | |
1091 } while (!_cm->has_aborted() && the_task->has_aborted()); | |
1092 } | |
1093 the_task->record_end_time(); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1094 guarantee(!the_task->has_aborted() || _cm->has_aborted(), "invariant"); |
342 | 1095 |
1096 ConcurrentGCThread::stsLeave(); | |
1097 | |
1098 double end_vtime = os::elapsedVTime(); | |
1099 _cm->update_accum_task_vtime(worker_i, end_vtime - start_vtime); | |
1100 } | |
1101 | |
1102 CMConcurrentMarkingTask(ConcurrentMark* cm, | |
1103 ConcurrentMarkThread* cmt) : | |
1104 AbstractGangTask("Concurrent Mark"), _cm(cm), _cmt(cmt) { } | |
1105 | |
1106 ~CMConcurrentMarkingTask() { } | |
1107 }; | |
1108 | |
1109 void ConcurrentMark::markFromRoots() { | |
1110 // we might be tempted to assert that: | |
1111 // assert(asynch == !SafepointSynchronize::is_at_safepoint(), | |
1112 // "inconsistent argument?"); | |
1113 // However that wouldn't be right, because it's possible that | |
1114 // a safepoint is indeed in progress as a younger generation | |
1115 // stop-the-world GC happens even as we mark in this generation. | |
1116 | |
1117 _restart_for_overflow = false; | |
1118 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1119 size_t active_workers = MAX2((size_t) 1, parallel_marking_threads()); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1120 set_phase(active_workers, true /* concurrent */); |
342 | 1121 |
1122 CMConcurrentMarkingTask markingTask(this, cmThread()); | |
1123 if (parallel_marking_threads() > 0) | |
1124 _parallel_workers->run_task(&markingTask); | |
1125 else | |
1126 markingTask.work(0); | |
1127 print_stats(); | |
1128 } | |
1129 | |
1130 void ConcurrentMark::checkpointRootsFinal(bool clear_all_soft_refs) { | |
1131 // world is stopped at this checkpoint | |
1132 assert(SafepointSynchronize::is_at_safepoint(), | |
1133 "world should be stopped"); | |
1134 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
1135 | |
1136 // If a full collection has happened, we shouldn't do this. | |
1137 if (has_aborted()) { | |
1138 g1h->set_marking_complete(); // So bitmap clearing isn't confused | |
1139 return; | |
1140 } | |
1141 | |
2125
7246a374a9f2
6458402: 3 jvmti tests fail with CMS and +ExplicitGCInvokesConcurrent
kamg
parents:
2037
diff
changeset
|
1142 SvcGCMarker sgcm(SvcGCMarker::OTHER); |
7246a374a9f2
6458402: 3 jvmti tests fail with CMS and +ExplicitGCInvokesConcurrent
kamg
parents:
2037
diff
changeset
|
1143 |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1144 if (VerifyDuringGC) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1145 HandleMark hm; // handle scope |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1146 gclog_or_tty->print(" VerifyDuringGC:(before)"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1147 Universe::heap()->prepare_for_verify(); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1148 Universe::verify(true, false, true); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1149 } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1150 |
342 | 1151 G1CollectorPolicy* g1p = g1h->g1_policy(); |
1152 g1p->record_concurrent_mark_remark_start(); | |
1153 | |
1154 double start = os::elapsedTime(); | |
1155 | |
1156 checkpointRootsFinalWork(); | |
1157 | |
1158 double mark_work_end = os::elapsedTime(); | |
1159 | |
1160 weakRefsWork(clear_all_soft_refs); | |
1161 | |
1162 if (has_overflown()) { | |
1163 // Oops. We overflowed. Restart concurrent marking. | |
1164 _restart_for_overflow = true; | |
1165 // Clear the flag. We do not need it any more. | |
1166 clear_has_overflown(); | |
1167 if (G1TraceMarkStackOverflow) | |
1168 gclog_or_tty->print_cr("\nRemark led to restart for overflow."); | |
1169 } else { | |
2149 | 1170 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); |
342 | 1171 // We're done with marking. |
1317
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
1172 // This is the end of the marking cycle, we're expected all |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
1173 // threads to have SATB queues with active set to true. |
2149 | 1174 satb_mq_set.set_active_all_threads(false, /* new active value */ |
1175 true /* expected_active */); | |
811 | 1176 |
1177 if (VerifyDuringGC) { | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1178 HandleMark hm; // handle scope |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1179 gclog_or_tty->print(" VerifyDuringGC:(after)"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1180 Universe::heap()->prepare_for_verify(); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1181 Universe::heap()->verify(/* allow_dirty */ true, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1182 /* silent */ false, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1183 /* use_prev_marking */ false); |
811 | 1184 } |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1185 assert(!restart_for_overflow(), "sanity"); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1186 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1187 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1188 // Reset the marking state if marking completed |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1189 if (!restart_for_overflow()) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1190 set_non_marking_state(); |
342 | 1191 } |
1192 | |
1193 #if VERIFY_OBJS_PROCESSED | |
1194 _scan_obj_cl.objs_processed = 0; | |
1195 ThreadLocalObjQueue::objs_enqueued = 0; | |
1196 #endif | |
1197 | |
1198 // Statistics | |
1199 double now = os::elapsedTime(); | |
1200 _remark_mark_times.add((mark_work_end - start) * 1000.0); | |
1201 _remark_weak_ref_times.add((now - mark_work_end) * 1000.0); | |
1202 _remark_times.add((now - start) * 1000.0); | |
1203 | |
1204 g1p->record_concurrent_mark_remark_end(); | |
1205 } | |
1206 | |
1207 #define CARD_BM_TEST_MODE 0 | |
1208 | |
1209 class CalcLiveObjectsClosure: public HeapRegionClosure { | |
1210 | |
1211 CMBitMapRO* _bm; | |
1212 ConcurrentMark* _cm; | |
1213 bool _changed; | |
1214 bool _yield; | |
1215 size_t _words_done; | |
1216 size_t _tot_live; | |
1217 size_t _tot_used; | |
1218 size_t _regions_done; | |
1219 double _start_vtime_sec; | |
1220 | |
1221 BitMap* _region_bm; | |
1222 BitMap* _card_bm; | |
1223 intptr_t _bottom_card_num; | |
1224 bool _final; | |
1225 | |
1226 void mark_card_num_range(intptr_t start_card_num, intptr_t last_card_num) { | |
1227 for (intptr_t i = start_card_num; i <= last_card_num; i++) { | |
1228 #if CARD_BM_TEST_MODE | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1229 guarantee(_card_bm->at(i - _bottom_card_num), "Should already be set."); |
342 | 1230 #else |
1231 _card_bm->par_at_put(i - _bottom_card_num, 1); | |
1232 #endif | |
1233 } | |
1234 } | |
1235 | |
1236 public: | |
1237 CalcLiveObjectsClosure(bool final, | |
1238 CMBitMapRO *bm, ConcurrentMark *cm, | |
936 | 1239 BitMap* region_bm, BitMap* card_bm) : |
342 | 1240 _bm(bm), _cm(cm), _changed(false), _yield(true), |
1241 _words_done(0), _tot_live(0), _tot_used(0), | |
936 | 1242 _region_bm(region_bm), _card_bm(card_bm),_final(final), |
342 | 1243 _regions_done(0), _start_vtime_sec(0.0) |
1244 { | |
1245 _bottom_card_num = | |
1246 intptr_t(uintptr_t(G1CollectedHeap::heap()->reserved_region().start()) >> | |
1247 CardTableModRefBS::card_shift); | |
1248 } | |
1249 | |
829 | 1250 // It takes a region that's not empty (i.e., it has at least one |
1251 // live object in it and sets its corresponding bit on the region | |
1252 // bitmap to 1. If the region is "starts humongous" it will also set | |
1253 // to 1 the bits on the region bitmap that correspond to its | |
1254 // associated "continues humongous" regions. | |
1255 void set_bit_for_region(HeapRegion* hr) { | |
1256 assert(!hr->continuesHumongous(), "should have filtered those out"); | |
1257 | |
1258 size_t index = hr->hrs_index(); | |
1259 if (!hr->startsHumongous()) { | |
1260 // Normal (non-humongous) case: just set the bit. | |
1261 _region_bm->par_at_put((BitMap::idx_t) index, true); | |
1262 } else { | |
1263 // Starts humongous case: calculate how many regions are part of | |
1264 // this humongous region and then set the bit range. It might | |
1265 // have been a bit more efficient to look at the object that | |
1266 // spans these humongous regions to calculate their number from | |
1267 // the object's size. However, it's a good idea to calculate | |
1268 // this based on the metadata itself, and not the region | |
1269 // contents, so that this code is not aware of what goes into | |
1270 // the humongous regions (in case this changes in the future). | |
1271 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
1272 size_t end_index = index + 1; | |
831 | 1273 while (end_index < g1h->n_regions()) { |
1274 HeapRegion* chr = g1h->region_at(end_index); | |
829 | 1275 if (!chr->continuesHumongous()) { |
1276 break; | |
1277 } | |
1278 end_index += 1; | |
1279 } | |
1280 _region_bm->par_at_put_range((BitMap::idx_t) index, | |
1281 (BitMap::idx_t) end_index, true); | |
1282 } | |
1283 } | |
1284 | |
342 | 1285 bool doHeapRegion(HeapRegion* hr) { |
1286 if (!_final && _regions_done == 0) | |
1287 _start_vtime_sec = os::elapsedVTime(); | |
1288 | |
639 | 1289 if (hr->continuesHumongous()) { |
829 | 1290 // We will ignore these here and process them when their |
1291 // associated "starts humongous" region is processed (see | |
1292 // set_bit_for_heap_region()). Note that we cannot rely on their | |
1293 // associated "starts humongous" region to have their bit set to | |
1294 // 1 since, due to the region chunking in the parallel region | |
1295 // iteration, a "continues humongous" region might be visited | |
1296 // before its associated "starts humongous". | |
639 | 1297 return false; |
1298 } | |
342 | 1299 |
1300 HeapWord* nextTop = hr->next_top_at_mark_start(); | |
1301 HeapWord* start = hr->top_at_conc_mark_count(); | |
1302 assert(hr->bottom() <= start && start <= hr->end() && | |
1303 hr->bottom() <= nextTop && nextTop <= hr->end() && | |
1304 start <= nextTop, | |
1305 "Preconditions."); | |
1306 // Otherwise, record the number of word's we'll examine. | |
1307 size_t words_done = (nextTop - start); | |
1308 // Find the first marked object at or after "start". | |
1309 start = _bm->getNextMarkedWordAddress(start, nextTop); | |
1310 size_t marked_bytes = 0; | |
1311 | |
1312 // Below, the term "card num" means the result of shifting an address | |
1313 // by the card shift -- address 0 corresponds to card number 0. One | |
1314 // must subtract the card num of the bottom of the heap to obtain a | |
1315 // card table index. | |
1316 // The first card num of the sequence of live cards currently being | |
1317 // constructed. -1 ==> no sequence. | |
1318 intptr_t start_card_num = -1; | |
1319 // The last card num of the sequence of live cards currently being | |
1320 // constructed. -1 ==> no sequence. | |
1321 intptr_t last_card_num = -1; | |
1322 | |
1323 while (start < nextTop) { | |
1324 if (_yield && _cm->do_yield_check()) { | |
1325 // We yielded. It might be for a full collection, in which case | |
1326 // all bets are off; terminate the traversal. | |
1327 if (_cm->has_aborted()) { | |
1328 _changed = false; | |
1329 return true; | |
1330 } else { | |
1331 // Otherwise, it might be a collection pause, and the region | |
1332 // we're looking at might be in the collection set. We'll | |
1333 // abandon this region. | |
1334 return false; | |
1335 } | |
1336 } | |
1337 oop obj = oop(start); | |
1338 int obj_sz = obj->size(); | |
1339 // The card num of the start of the current object. | |
1340 intptr_t obj_card_num = | |
1341 intptr_t(uintptr_t(start) >> CardTableModRefBS::card_shift); | |
1342 | |
1343 HeapWord* obj_last = start + obj_sz - 1; | |
1344 intptr_t obj_last_card_num = | |
1345 intptr_t(uintptr_t(obj_last) >> CardTableModRefBS::card_shift); | |
1346 | |
1347 if (obj_card_num != last_card_num) { | |
1348 if (start_card_num == -1) { | |
1349 assert(last_card_num == -1, "Both or neither."); | |
1350 start_card_num = obj_card_num; | |
1351 } else { | |
1352 assert(last_card_num != -1, "Both or neither."); | |
1353 assert(obj_card_num >= last_card_num, "Inv"); | |
1354 if ((obj_card_num - last_card_num) > 1) { | |
1355 // Mark the last run, and start a new one. | |
1356 mark_card_num_range(start_card_num, last_card_num); | |
1357 start_card_num = obj_card_num; | |
1358 } | |
1359 } | |
1360 #if CARD_BM_TEST_MODE | |
1361 /* | |
1362 gclog_or_tty->print_cr("Setting bits from %d/%d.", | |
1363 obj_card_num - _bottom_card_num, | |
1364 obj_last_card_num - _bottom_card_num); | |
1365 */ | |
1366 for (intptr_t j = obj_card_num; j <= obj_last_card_num; j++) { | |
1367 _card_bm->par_at_put(j - _bottom_card_num, 1); | |
1368 } | |
1369 #endif | |
1370 } | |
1371 // In any case, we set the last card num. | |
1372 last_card_num = obj_last_card_num; | |
1373 | |
1030
dfdaf65c3423
6858886: G1: guarantee(_next_marked_bytes <= used(),"invariant") at heapRegion.hpp:359
apetrusenko
parents:
1023
diff
changeset
|
1374 marked_bytes += (size_t)obj_sz * HeapWordSize; |
342 | 1375 // Find the next marked object after this one. |
1376 start = _bm->getNextMarkedWordAddress(start + 1, nextTop); | |
1377 _changed = true; | |
1378 } | |
1379 // Handle the last range, if any. | |
1380 if (start_card_num != -1) | |
1381 mark_card_num_range(start_card_num, last_card_num); | |
1382 if (_final) { | |
1383 // Mark the allocated-since-marking portion... | |
1384 HeapWord* tp = hr->top(); | |
1385 if (nextTop < tp) { | |
1386 start_card_num = | |
1387 intptr_t(uintptr_t(nextTop) >> CardTableModRefBS::card_shift); | |
1388 last_card_num = | |
1389 intptr_t(uintptr_t(tp) >> CardTableModRefBS::card_shift); | |
1390 mark_card_num_range(start_card_num, last_card_num); | |
1391 // This definitely means the region has live objects. | |
829 | 1392 set_bit_for_region(hr); |
342 | 1393 } |
1394 } | |
1395 | |
1396 hr->add_to_marked_bytes(marked_bytes); | |
1397 // Update the live region bitmap. | |
1398 if (marked_bytes > 0) { | |
829 | 1399 set_bit_for_region(hr); |
342 | 1400 } |
1401 hr->set_top_at_conc_mark_count(nextTop); | |
1402 _tot_live += hr->next_live_bytes(); | |
1403 _tot_used += hr->used(); | |
1404 _words_done = words_done; | |
1405 | |
1406 if (!_final) { | |
1407 ++_regions_done; | |
1408 if (_regions_done % 10 == 0) { | |
1409 double end_vtime_sec = os::elapsedVTime(); | |
1410 double elapsed_vtime_sec = end_vtime_sec - _start_vtime_sec; | |
1411 if (elapsed_vtime_sec > (10.0 / 1000.0)) { | |
1412 jlong sleep_time_ms = | |
1413 (jlong) (elapsed_vtime_sec * _cm->cleanup_sleep_factor() * 1000.0); | |
1414 os::sleep(Thread::current(), sleep_time_ms, false); | |
1415 _start_vtime_sec = end_vtime_sec; | |
1416 } | |
1417 } | |
1418 } | |
1419 | |
1420 return false; | |
1421 } | |
1422 | |
1423 bool changed() { return _changed; } | |
1424 void reset() { _changed = false; _words_done = 0; } | |
1425 void no_yield() { _yield = false; } | |
1426 size_t words_done() { return _words_done; } | |
1427 size_t tot_live() { return _tot_live; } | |
1428 size_t tot_used() { return _tot_used; } | |
1429 }; | |
1430 | |
1431 | |
1432 void ConcurrentMark::calcDesiredRegions() { | |
1433 _region_bm.clear(); | |
1434 _card_bm.clear(); | |
1435 CalcLiveObjectsClosure calccl(false /*final*/, | |
1436 nextMarkBitMap(), this, | |
936 | 1437 &_region_bm, &_card_bm); |
342 | 1438 G1CollectedHeap *g1h = G1CollectedHeap::heap(); |
1439 g1h->heap_region_iterate(&calccl); | |
1440 | |
1441 do { | |
1442 calccl.reset(); | |
1443 g1h->heap_region_iterate(&calccl); | |
1444 } while (calccl.changed()); | |
1445 } | |
1446 | |
1447 class G1ParFinalCountTask: public AbstractGangTask { | |
1448 protected: | |
1449 G1CollectedHeap* _g1h; | |
1450 CMBitMap* _bm; | |
1451 size_t _n_workers; | |
1452 size_t *_live_bytes; | |
1453 size_t *_used_bytes; | |
1454 BitMap* _region_bm; | |
1455 BitMap* _card_bm; | |
1456 public: | |
1457 G1ParFinalCountTask(G1CollectedHeap* g1h, CMBitMap* bm, | |
1458 BitMap* region_bm, BitMap* card_bm) : | |
1459 AbstractGangTask("G1 final counting"), _g1h(g1h), | |
1460 _bm(bm), _region_bm(region_bm), _card_bm(card_bm) | |
1461 { | |
1462 if (ParallelGCThreads > 0) | |
1463 _n_workers = _g1h->workers()->total_workers(); | |
1464 else | |
1465 _n_workers = 1; | |
1466 _live_bytes = NEW_C_HEAP_ARRAY(size_t, _n_workers); | |
1467 _used_bytes = NEW_C_HEAP_ARRAY(size_t, _n_workers); | |
1468 } | |
1469 | |
1470 ~G1ParFinalCountTask() { | |
1471 FREE_C_HEAP_ARRAY(size_t, _live_bytes); | |
1472 FREE_C_HEAP_ARRAY(size_t, _used_bytes); | |
1473 } | |
1474 | |
1475 void work(int i) { | |
1476 CalcLiveObjectsClosure calccl(true /*final*/, | |
1477 _bm, _g1h->concurrent_mark(), | |
936 | 1478 _region_bm, _card_bm); |
342 | 1479 calccl.no_yield(); |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1480 if (G1CollectedHeap::use_parallel_gc_threads()) { |
355 | 1481 _g1h->heap_region_par_iterate_chunked(&calccl, i, |
1482 HeapRegion::FinalCountClaimValue); | |
342 | 1483 } else { |
1484 _g1h->heap_region_iterate(&calccl); | |
1485 } | |
1486 assert(calccl.complete(), "Shouldn't have yielded!"); | |
1487 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
1488 assert((size_t) i < _n_workers, "invariant"); |
342 | 1489 _live_bytes[i] = calccl.tot_live(); |
1490 _used_bytes[i] = calccl.tot_used(); | |
1491 } | |
1492 size_t live_bytes() { | |
1493 size_t live_bytes = 0; | |
1494 for (size_t i = 0; i < _n_workers; ++i) | |
1495 live_bytes += _live_bytes[i]; | |
1496 return live_bytes; | |
1497 } | |
1498 size_t used_bytes() { | |
1499 size_t used_bytes = 0; | |
1500 for (size_t i = 0; i < _n_workers; ++i) | |
1501 used_bytes += _used_bytes[i]; | |
1502 return used_bytes; | |
1503 } | |
1504 }; | |
1505 | |
1506 class G1ParNoteEndTask; | |
1507 | |
1508 class G1NoteEndOfConcMarkClosure : public HeapRegionClosure { | |
1509 G1CollectedHeap* _g1; | |
1510 int _worker_num; | |
1511 size_t _max_live_bytes; | |
1512 size_t _regions_claimed; | |
1513 size_t _freed_bytes; | |
2173 | 1514 FreeRegionList* _local_cleanup_list; |
1515 HumongousRegionSet* _humongous_proxy_set; | |
1516 HRRSCleanupTask* _hrrs_cleanup_task; | |
342 | 1517 double _claimed_region_time; |
1518 double _max_region_time; | |
1519 | |
1520 public: | |
1521 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1, | |
2173 | 1522 int worker_num, |
1523 FreeRegionList* local_cleanup_list, | |
1524 HumongousRegionSet* humongous_proxy_set, | |
1525 HRRSCleanupTask* hrrs_cleanup_task); | |
342 | 1526 size_t freed_bytes() { return _freed_bytes; } |
1527 | |
1528 bool doHeapRegion(HeapRegion *r); | |
1529 | |
1530 size_t max_live_bytes() { return _max_live_bytes; } | |
1531 size_t regions_claimed() { return _regions_claimed; } | |
1532 double claimed_region_time_sec() { return _claimed_region_time; } | |
1533 double max_region_time_sec() { return _max_region_time; } | |
1534 }; | |
1535 | |
1536 class G1ParNoteEndTask: public AbstractGangTask { | |
1537 friend class G1NoteEndOfConcMarkClosure; | |
2152 | 1538 |
342 | 1539 protected: |
1540 G1CollectedHeap* _g1h; | |
1541 size_t _max_live_bytes; | |
1542 size_t _freed_bytes; | |
2152 | 1543 FreeRegionList* _cleanup_list; |
1544 | |
342 | 1545 public: |
1546 G1ParNoteEndTask(G1CollectedHeap* g1h, | |
2152 | 1547 FreeRegionList* cleanup_list) : |
342 | 1548 AbstractGangTask("G1 note end"), _g1h(g1h), |
2152 | 1549 _max_live_bytes(0), _freed_bytes(0), _cleanup_list(cleanup_list) { } |
342 | 1550 |
1551 void work(int i) { | |
1552 double start = os::elapsedTime(); | |
2173 | 1553 FreeRegionList local_cleanup_list("Local Cleanup List"); |
1554 HumongousRegionSet humongous_proxy_set("Local Cleanup Humongous Proxy Set"); | |
1555 HRRSCleanupTask hrrs_cleanup_task; | |
1556 G1NoteEndOfConcMarkClosure g1_note_end(_g1h, i, &local_cleanup_list, | |
1557 &humongous_proxy_set, | |
1558 &hrrs_cleanup_task); | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1559 if (G1CollectedHeap::use_parallel_gc_threads()) { |
355 | 1560 _g1h->heap_region_par_iterate_chunked(&g1_note_end, i, |
1561 HeapRegion::NoteEndClaimValue); | |
342 | 1562 } else { |
1563 _g1h->heap_region_iterate(&g1_note_end); | |
1564 } | |
1565 assert(g1_note_end.complete(), "Shouldn't have yielded!"); | |
1566 | |
2152 | 1567 // Now update the lists |
1568 _g1h->update_sets_after_freeing_regions(g1_note_end.freed_bytes(), | |
1569 NULL /* free_list */, | |
2173 | 1570 &humongous_proxy_set, |
2152 | 1571 true /* par */); |
342 | 1572 { |
1573 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); | |
1574 _max_live_bytes += g1_note_end.max_live_bytes(); | |
1575 _freed_bytes += g1_note_end.freed_bytes(); | |
2152 | 1576 |
2173 | 1577 _cleanup_list->add_as_tail(&local_cleanup_list); |
1578 assert(local_cleanup_list.is_empty(), "post-condition"); | |
1579 | |
1580 HeapRegionRemSet::finish_cleanup_task(&hrrs_cleanup_task); | |
342 | 1581 } |
1582 double end = os::elapsedTime(); | |
1583 if (G1PrintParCleanupStats) { | |
1584 gclog_or_tty->print(" Worker thread %d [%8.3f..%8.3f = %8.3f ms] " | |
1585 "claimed %d regions (tot = %8.3f ms, max = %8.3f ms).\n", | |
1586 i, start, end, (end-start)*1000.0, | |
1587 g1_note_end.regions_claimed(), | |
1588 g1_note_end.claimed_region_time_sec()*1000.0, | |
1589 g1_note_end.max_region_time_sec()*1000.0); | |
1590 } | |
1591 } | |
1592 size_t max_live_bytes() { return _max_live_bytes; } | |
1593 size_t freed_bytes() { return _freed_bytes; } | |
1594 }; | |
1595 | |
1596 class G1ParScrubRemSetTask: public AbstractGangTask { | |
1597 protected: | |
1598 G1RemSet* _g1rs; | |
1599 BitMap* _region_bm; | |
1600 BitMap* _card_bm; | |
1601 public: | |
1602 G1ParScrubRemSetTask(G1CollectedHeap* g1h, | |
1603 BitMap* region_bm, BitMap* card_bm) : | |
1604 AbstractGangTask("G1 ScrubRS"), _g1rs(g1h->g1_rem_set()), | |
1605 _region_bm(region_bm), _card_bm(card_bm) | |
1606 {} | |
1607 | |
1608 void work(int i) { | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1609 if (G1CollectedHeap::use_parallel_gc_threads()) { |
355 | 1610 _g1rs->scrub_par(_region_bm, _card_bm, i, |
1611 HeapRegion::ScrubRemSetClaimValue); | |
342 | 1612 } else { |
1613 _g1rs->scrub(_region_bm, _card_bm); | |
1614 } | |
1615 } | |
1616 | |
1617 }; | |
1618 | |
1619 G1NoteEndOfConcMarkClosure:: | |
1620 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1, | |
2173 | 1621 int worker_num, |
1622 FreeRegionList* local_cleanup_list, | |
1623 HumongousRegionSet* humongous_proxy_set, | |
1624 HRRSCleanupTask* hrrs_cleanup_task) | |
342 | 1625 : _g1(g1), _worker_num(worker_num), |
1626 _max_live_bytes(0), _regions_claimed(0), | |
2152 | 1627 _freed_bytes(0), |
342 | 1628 _claimed_region_time(0.0), _max_region_time(0.0), |
2173 | 1629 _local_cleanup_list(local_cleanup_list), |
1630 _humongous_proxy_set(humongous_proxy_set), | |
1631 _hrrs_cleanup_task(hrrs_cleanup_task) { } | |
2152 | 1632 |
1633 bool G1NoteEndOfConcMarkClosure::doHeapRegion(HeapRegion *hr) { | |
342 | 1634 // We use a claim value of zero here because all regions |
1635 // were claimed with value 1 in the FinalCount task. | |
2152 | 1636 hr->reset_gc_time_stamp(); |
1637 if (!hr->continuesHumongous()) { | |
342 | 1638 double start = os::elapsedTime(); |
1639 _regions_claimed++; | |
2152 | 1640 hr->note_end_of_marking(); |
1641 _max_live_bytes += hr->max_live_bytes(); | |
2173 | 1642 _g1->free_region_if_empty(hr, |
1643 &_freed_bytes, | |
1644 _local_cleanup_list, | |
1645 _humongous_proxy_set, | |
1646 _hrrs_cleanup_task, | |
1647 true /* par */); | |
342 | 1648 double region_time = (os::elapsedTime() - start); |
1649 _claimed_region_time += region_time; | |
1650 if (region_time > _max_region_time) _max_region_time = region_time; | |
1651 } | |
1652 return false; | |
1653 } | |
1654 | |
1655 void ConcurrentMark::cleanup() { | |
1656 // world is stopped at this checkpoint | |
1657 assert(SafepointSynchronize::is_at_safepoint(), | |
1658 "world should be stopped"); | |
1659 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
1660 | |
1661 // If a full collection has happened, we shouldn't do this. | |
1662 if (has_aborted()) { | |
1663 g1h->set_marking_complete(); // So bitmap clearing isn't confused | |
1664 return; | |
1665 } | |
1666 | |
2152 | 1667 g1h->verify_region_sets_optional(); |
1668 | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1669 if (VerifyDuringGC) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1670 HandleMark hm; // handle scope |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1671 gclog_or_tty->print(" VerifyDuringGC:(before)"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1672 Universe::heap()->prepare_for_verify(); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1673 Universe::verify(/* allow dirty */ true, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1674 /* silent */ false, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1675 /* prev marking */ true); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1676 } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1677 |
342 | 1678 G1CollectorPolicy* g1p = G1CollectedHeap::heap()->g1_policy(); |
1679 g1p->record_concurrent_mark_cleanup_start(); | |
1680 | |
1681 double start = os::elapsedTime(); | |
1682 | |
2173 | 1683 HeapRegionRemSet::reset_for_cleanup_tasks(); |
1684 | |
342 | 1685 // Do counting once more with the world stopped for good measure. |
1686 G1ParFinalCountTask g1_par_count_task(g1h, nextMarkBitMap(), | |
1687 &_region_bm, &_card_bm); | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1688 if (G1CollectedHeap::use_parallel_gc_threads()) { |
355 | 1689 assert(g1h->check_heap_region_claim_values( |
1690 HeapRegion::InitialClaimValue), | |
1691 "sanity check"); | |
1692 | |
342 | 1693 int n_workers = g1h->workers()->total_workers(); |
1694 g1h->set_par_threads(n_workers); | |
1695 g1h->workers()->run_task(&g1_par_count_task); | |
1696 g1h->set_par_threads(0); | |
355 | 1697 |
1698 assert(g1h->check_heap_region_claim_values( | |
1699 HeapRegion::FinalCountClaimValue), | |
1700 "sanity check"); | |
342 | 1701 } else { |
1702 g1_par_count_task.work(0); | |
1703 } | |
1704 | |
1705 size_t known_garbage_bytes = | |
1706 g1_par_count_task.used_bytes() - g1_par_count_task.live_bytes(); | |
1707 #if 0 | |
1708 gclog_or_tty->print_cr("used %1.2lf, live %1.2lf, garbage %1.2lf", | |
1709 (double) g1_par_count_task.used_bytes() / (double) (1024 * 1024), | |
1710 (double) g1_par_count_task.live_bytes() / (double) (1024 * 1024), | |
1711 (double) known_garbage_bytes / (double) (1024 * 1024)); | |
1712 #endif // 0 | |
1713 g1p->set_known_garbage_bytes(known_garbage_bytes); | |
1714 | |
1715 size_t start_used_bytes = g1h->used(); | |
1716 _at_least_one_mark_complete = true; | |
1717 g1h->set_marking_complete(); | |
1718 | |
1719 double count_end = os::elapsedTime(); | |
1720 double this_final_counting_time = (count_end - start); | |
1721 if (G1PrintParCleanupStats) { | |
1722 gclog_or_tty->print_cr("Cleanup:"); | |
1723 gclog_or_tty->print_cr(" Finalize counting: %8.3f ms", | |
1724 this_final_counting_time*1000.0); | |
1725 } | |
1726 _total_counting_time += this_final_counting_time; | |
1727 | |
2435
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
1728 if (G1PrintRegionLivenessInfo) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
1729 G1PrintRegionLivenessInfoClosure cl(gclog_or_tty, "Post-Marking"); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
1730 _g1h->heap_region_iterate(&cl); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
1731 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
1732 |
342 | 1733 // Install newly created mark bitMap as "prev". |
1734 swapMarkBitMaps(); | |
1735 | |
1736 g1h->reset_gc_time_stamp(); | |
1737 | |
1738 // Note end of marking in all heap regions. | |
1739 double note_end_start = os::elapsedTime(); | |
2152 | 1740 G1ParNoteEndTask g1_par_note_end_task(g1h, &_cleanup_list); |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1741 if (G1CollectedHeap::use_parallel_gc_threads()) { |
342 | 1742 int n_workers = g1h->workers()->total_workers(); |
1743 g1h->set_par_threads(n_workers); | |
1744 g1h->workers()->run_task(&g1_par_note_end_task); | |
1745 g1h->set_par_threads(0); | |
355 | 1746 |
1747 assert(g1h->check_heap_region_claim_values(HeapRegion::NoteEndClaimValue), | |
1748 "sanity check"); | |
342 | 1749 } else { |
1750 g1_par_note_end_task.work(0); | |
1751 } | |
2152 | 1752 |
1753 if (!cleanup_list_is_empty()) { | |
1754 // The cleanup list is not empty, so we'll have to process it | |
1755 // concurrently. Notify anyone else that might be wanting free | |
1756 // regions that there will be more free regions coming soon. | |
1757 g1h->set_free_regions_coming(); | |
1758 } | |
342 | 1759 double note_end_end = os::elapsedTime(); |
1760 if (G1PrintParCleanupStats) { | |
1761 gclog_or_tty->print_cr(" note end of marking: %8.3f ms.", | |
1762 (note_end_end - note_end_start)*1000.0); | |
1763 } | |
1764 | |
355 | 1765 |
342 | 1766 // call below, since it affects the metric by which we sort the heap |
1767 // regions. | |
1768 if (G1ScrubRemSets) { | |
1769 double rs_scrub_start = os::elapsedTime(); | |
1770 G1ParScrubRemSetTask g1_par_scrub_rs_task(g1h, &_region_bm, &_card_bm); | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
1771 if (G1CollectedHeap::use_parallel_gc_threads()) { |
342 | 1772 int n_workers = g1h->workers()->total_workers(); |
1773 g1h->set_par_threads(n_workers); | |
1774 g1h->workers()->run_task(&g1_par_scrub_rs_task); | |
1775 g1h->set_par_threads(0); | |
355 | 1776 |
1777 assert(g1h->check_heap_region_claim_values( | |
1778 HeapRegion::ScrubRemSetClaimValue), | |
1779 "sanity check"); | |
342 | 1780 } else { |
1781 g1_par_scrub_rs_task.work(0); | |
1782 } | |
1783 | |
1784 double rs_scrub_end = os::elapsedTime(); | |
1785 double this_rs_scrub_time = (rs_scrub_end - rs_scrub_start); | |
1786 _total_rs_scrub_time += this_rs_scrub_time; | |
1787 } | |
1788 | |
1789 // this will also free any regions totally full of garbage objects, | |
1790 // and sort the regions. | |
1791 g1h->g1_policy()->record_concurrent_mark_cleanup_end( | |
1792 g1_par_note_end_task.freed_bytes(), | |
1793 g1_par_note_end_task.max_live_bytes()); | |
1794 | |
1795 // Statistics. | |
1796 double end = os::elapsedTime(); | |
1797 _cleanup_times.add((end - start) * 1000.0); | |
1798 | |
1799 // G1CollectedHeap::heap()->print(); | |
1800 // gclog_or_tty->print_cr("HEAP GC TIME STAMP : %d", | |
1801 // G1CollectedHeap::heap()->get_gc_time_stamp()); | |
1802 | |
1803 if (PrintGC || PrintGCDetails) { | |
1804 g1h->print_size_transition(gclog_or_tty, | |
1805 start_used_bytes, | |
1806 g1h->used(), | |
1807 g1h->capacity()); | |
1808 } | |
1809 | |
1810 size_t cleaned_up_bytes = start_used_bytes - g1h->used(); | |
1811 g1p->decrease_known_garbage_bytes(cleaned_up_bytes); | |
1812 | |
1813 // We need to make this be a "collection" so any collection pause that | |
1814 // races with it goes around and waits for completeCleanup to finish. | |
1815 g1h->increment_total_collections(); | |
1816 | |
751 | 1817 if (VerifyDuringGC) { |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1818 HandleMark hm; // handle scope |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1819 gclog_or_tty->print(" VerifyDuringGC:(after)"); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1820 Universe::heap()->prepare_for_verify(); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1821 Universe::verify(/* allow dirty */ true, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1822 /* silent */ false, |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1823 /* prev marking */ true); |
342 | 1824 } |
2152 | 1825 |
1826 g1h->verify_region_sets_optional(); | |
342 | 1827 } |
1828 | |
1829 void ConcurrentMark::completeCleanup() { | |
1830 if (has_aborted()) return; | |
1831 | |
2152 | 1832 G1CollectedHeap* g1h = G1CollectedHeap::heap(); |
1833 | |
1834 _cleanup_list.verify_optional(); | |
2361 | 1835 FreeRegionList tmp_free_list("Tmp Free List"); |
2152 | 1836 |
1837 if (G1ConcRegionFreeingVerbose) { | |
1838 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : " | |
1839 "cleanup list has "SIZE_FORMAT" entries", | |
1840 _cleanup_list.length()); | |
1841 } | |
1842 | |
1843 // Noone else should be accessing the _cleanup_list at this point, | |
1844 // so it's not necessary to take any locks | |
1845 while (!_cleanup_list.is_empty()) { | |
1846 HeapRegion* hr = _cleanup_list.remove_head(); | |
1847 assert(hr != NULL, "the list was not empty"); | |
1848 hr->rem_set()->clear(); | |
2361 | 1849 tmp_free_list.add_as_tail(hr); |
2152 | 1850 |
1851 // Instead of adding one region at a time to the secondary_free_list, | |
1852 // we accumulate them in the local list and move them a few at a | |
1853 // time. This also cuts down on the number of notify_all() calls | |
1854 // we do during this process. We'll also append the local list when | |
1855 // _cleanup_list is empty (which means we just removed the last | |
1856 // region from the _cleanup_list). | |
2361 | 1857 if ((tmp_free_list.length() % G1SecondaryFreeListAppendLength == 0) || |
2152 | 1858 _cleanup_list.is_empty()) { |
1859 if (G1ConcRegionFreeingVerbose) { | |
1860 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : " | |
1861 "appending "SIZE_FORMAT" entries to the " | |
1862 "secondary_free_list, clean list still has " | |
1863 SIZE_FORMAT" entries", | |
2361 | 1864 tmp_free_list.length(), |
2152 | 1865 _cleanup_list.length()); |
342 | 1866 } |
2152 | 1867 |
1868 { | |
1869 MutexLockerEx x(SecondaryFreeList_lock, Mutex::_no_safepoint_check_flag); | |
2361 | 1870 g1h->secondary_free_list_add_as_tail(&tmp_free_list); |
2152 | 1871 SecondaryFreeList_lock->notify_all(); |
1872 } | |
1873 | |
1874 if (G1StressConcRegionFreeing) { | |
1875 for (uintx i = 0; i < G1StressConcRegionFreeingDelayMillis; ++i) { | |
1876 os::sleep(Thread::current(), (jlong) 1, false); | |
1877 } | |
1878 } | |
342 | 1879 } |
1880 } | |
2361 | 1881 assert(tmp_free_list.is_empty(), "post-condition"); |
342 | 1882 } |
1883 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1884 // Support closures for reference procssing in G1 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1885 |
2037
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
1886 bool G1CMIsAliveClosure::do_object_b(oop obj) { |
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
1887 HeapWord* addr = (HeapWord*)obj; |
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
1888 return addr != NULL && |
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
1889 (!_g1->is_in_g1_reserved(addr) || !_g1->is_obj_ill(obj)); |
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
1890 } |
342 | 1891 |
1892 class G1CMKeepAliveClosure: public OopClosure { | |
1893 G1CollectedHeap* _g1; | |
1894 ConcurrentMark* _cm; | |
1895 CMBitMap* _bitMap; | |
1896 public: | |
1897 G1CMKeepAliveClosure(G1CollectedHeap* g1, ConcurrentMark* cm, | |
1898 CMBitMap* bitMap) : | |
1899 _g1(g1), _cm(cm), | |
1900 _bitMap(bitMap) {} | |
1901 | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1902 virtual void do_oop(narrowOop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1903 virtual void do_oop( oop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1904 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
1905 template <class T> void do_oop_work(T* p) { |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1906 oop obj = oopDesc::load_decode_heap_oop(p); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1907 HeapWord* addr = (HeapWord*)obj; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1908 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1909 if (_cm->verbose_high()) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1910 gclog_or_tty->print_cr("\t[0] we're looking at location " |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1911 "*"PTR_FORMAT" = "PTR_FORMAT, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1912 p, (void*) obj); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1913 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1914 if (_g1->is_in_g1_reserved(addr) && _g1->is_obj_ill(obj)) { |
342 | 1915 _bitMap->mark(addr); |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1916 _cm->mark_stack_push(obj); |
342 | 1917 } |
1918 } | |
1919 }; | |
1920 | |
1921 class G1CMDrainMarkingStackClosure: public VoidClosure { | |
1922 CMMarkStack* _markStack; | |
1923 CMBitMap* _bitMap; | |
1924 G1CMKeepAliveClosure* _oopClosure; | |
1925 public: | |
1926 G1CMDrainMarkingStackClosure(CMBitMap* bitMap, CMMarkStack* markStack, | |
1927 G1CMKeepAliveClosure* oopClosure) : | |
1928 _bitMap(bitMap), | |
1929 _markStack(markStack), | |
1930 _oopClosure(oopClosure) | |
1931 {} | |
1932 | |
1933 void do_void() { | |
1934 _markStack->drain((OopClosure*)_oopClosure, _bitMap, false); | |
1935 } | |
1936 }; | |
1937 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1938 // 'Keep Alive' closure used by parallel reference processing. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1939 // An instance of this closure is used in the parallel reference processing |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1940 // code rather than an instance of G1CMKeepAliveClosure. We could have used |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1941 // the G1CMKeepAliveClosure as it is MT-safe. Also reference objects are |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1942 // placed on to discovered ref lists once so we can mark and push with no |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1943 // need to check whether the object has already been marked. Using the |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1944 // G1CMKeepAliveClosure would mean, however, having all the worker threads |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1945 // operating on the global mark stack. This means that an individual |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1946 // worker would be doing lock-free pushes while it processes its own |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1947 // discovered ref list followed by drain call. If the discovered ref lists |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1948 // are unbalanced then this could cause interference with the other |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1949 // workers. Using a CMTask (and its embedded local data structures) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1950 // avoids that potential interference. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1951 class G1CMParKeepAliveAndDrainClosure: public OopClosure { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1952 ConcurrentMark* _cm; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1953 CMTask* _task; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1954 CMBitMap* _bitMap; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1955 int _ref_counter_limit; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1956 int _ref_counter; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1957 public: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1958 G1CMParKeepAliveAndDrainClosure(ConcurrentMark* cm, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1959 CMTask* task, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1960 CMBitMap* bitMap) : |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1961 _cm(cm), _task(task), _bitMap(bitMap), |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1962 _ref_counter_limit(G1RefProcDrainInterval) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1963 { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1964 assert(_ref_counter_limit > 0, "sanity"); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1965 _ref_counter = _ref_counter_limit; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1966 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1967 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1968 virtual void do_oop(narrowOop* p) { do_oop_work(p); } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1969 virtual void do_oop( oop* p) { do_oop_work(p); } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1970 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1971 template <class T> void do_oop_work(T* p) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1972 if (!_cm->has_overflown()) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1973 oop obj = oopDesc::load_decode_heap_oop(p); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1974 if (_cm->verbose_high()) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1975 gclog_or_tty->print_cr("\t[%d] we're looking at location " |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1976 "*"PTR_FORMAT" = "PTR_FORMAT, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1977 _task->task_id(), p, (void*) obj); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1978 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1979 _task->deal_with_reference(obj); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1980 _ref_counter--; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1981 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1982 if (_ref_counter == 0) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1983 // We have dealt with _ref_counter_limit references, pushing them and objects |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1984 // reachable from them on to the local stack (and possibly the global stack). |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1985 // Call do_marking_step() to process these entries. We call the routine in a |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1986 // loop, which we'll exit if there's nothing more to do (i.e. we're done |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1987 // with the entries that we've pushed as a result of the deal_with_reference |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1988 // calls above) or we overflow. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1989 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted() flag |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1990 // while there may still be some work to do. (See the comment at the |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1991 // beginning of CMTask::do_marking_step() for those conditions - one of which |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1992 // is reaching the specified time target.) It is only when |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1993 // CMTask::do_marking_step() returns without setting the has_aborted() flag |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1994 // that the marking has completed. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1995 do { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1996 double mark_step_duration_ms = G1ConcMarkStepDurationMillis; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1997 _task->do_marking_step(mark_step_duration_ms, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1998 false /* do_stealing */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
1999 false /* do_termination */); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2000 } while (_task->has_aborted() && !_cm->has_overflown()); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2001 _ref_counter = _ref_counter_limit; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2002 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2003 } else { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2004 if (_cm->verbose_high()) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2005 gclog_or_tty->print_cr("\t[%d] CM Overflow", _task->task_id()); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2006 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2007 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2008 }; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2009 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2010 class G1CMParDrainMarkingStackClosure: public VoidClosure { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2011 ConcurrentMark* _cm; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2012 CMTask* _task; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2013 public: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2014 G1CMParDrainMarkingStackClosure(ConcurrentMark* cm, CMTask* task) : |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2015 _cm(cm), _task(task) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2016 {} |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2017 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2018 void do_void() { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2019 do { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2020 if (_cm->verbose_high()) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2021 gclog_or_tty->print_cr("\t[%d] Drain: Calling do marking_step", _task->task_id()); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2022 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2023 // We call CMTask::do_marking_step() to completely drain the local and |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2024 // global marking stacks. The routine is called in a loop, which we'll |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2025 // exit if there's nothing more to do (i.e. we'completely drained the |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2026 // entries that were pushed as a result of applying the |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2027 // G1CMParKeepAliveAndDrainClosure to the entries on the discovered ref |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2028 // lists above) or we overflow the global marking stack. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2029 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted() flag |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2030 // while there may still be some work to do. (See the comment at the |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2031 // beginning of CMTask::do_marking_step() for those conditions - one of which |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2032 // is reaching the specified time target.) It is only when |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2033 // CMTask::do_marking_step() returns without setting the has_aborted() flag |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2034 // that the marking has completed. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2035 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2036 _task->do_marking_step(1000000000.0 /* something very large */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2037 true /* do_stealing */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2038 true /* do_termination */); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2039 } while (_task->has_aborted() && !_cm->has_overflown()); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2040 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2041 }; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2042 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2043 // Implementation of AbstractRefProcTaskExecutor for G1 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2044 class G1RefProcTaskExecutor: public AbstractRefProcTaskExecutor { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2045 private: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2046 G1CollectedHeap* _g1h; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2047 ConcurrentMark* _cm; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2048 CMBitMap* _bitmap; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2049 WorkGang* _workers; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2050 int _active_workers; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2051 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2052 public: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2053 G1RefProcTaskExecutor(G1CollectedHeap* g1h, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2054 ConcurrentMark* cm, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2055 CMBitMap* bitmap, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2056 WorkGang* workers, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2057 int n_workers) : |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2058 _g1h(g1h), _cm(cm), _bitmap(bitmap), |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2059 _workers(workers), _active_workers(n_workers) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2060 { } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2061 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2062 // Executes the given task using concurrent marking worker threads. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2063 virtual void execute(ProcessTask& task); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2064 virtual void execute(EnqueueTask& task); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2065 }; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2066 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2067 class G1RefProcTaskProxy: public AbstractGangTask { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2068 typedef AbstractRefProcTaskExecutor::ProcessTask ProcessTask; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2069 ProcessTask& _proc_task; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2070 G1CollectedHeap* _g1h; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2071 ConcurrentMark* _cm; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2072 CMBitMap* _bitmap; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2073 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2074 public: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2075 G1RefProcTaskProxy(ProcessTask& proc_task, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2076 G1CollectedHeap* g1h, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2077 ConcurrentMark* cm, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2078 CMBitMap* bitmap) : |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2079 AbstractGangTask("Process reference objects in parallel"), |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2080 _proc_task(proc_task), _g1h(g1h), _cm(cm), _bitmap(bitmap) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2081 {} |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2082 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2083 virtual void work(int i) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2084 CMTask* marking_task = _cm->task(i); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2085 G1CMIsAliveClosure g1_is_alive(_g1h); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2086 G1CMParKeepAliveAndDrainClosure g1_par_keep_alive(_cm, marking_task, _bitmap); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2087 G1CMParDrainMarkingStackClosure g1_par_drain(_cm, marking_task); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2088 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2089 _proc_task.work(i, g1_is_alive, g1_par_keep_alive, g1_par_drain); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2090 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2091 }; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2092 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2093 void G1RefProcTaskExecutor::execute(ProcessTask& proc_task) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2094 assert(_workers != NULL, "Need parallel worker threads."); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2095 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2096 G1RefProcTaskProxy proc_task_proxy(proc_task, _g1h, _cm, _bitmap); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2097 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2098 // We need to reset the phase for each task execution so that |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2099 // the termination protocol of CMTask::do_marking_step works. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2100 _cm->set_phase(_active_workers, false /* concurrent */); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2101 _g1h->set_par_threads(_active_workers); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2102 _workers->run_task(&proc_task_proxy); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2103 _g1h->set_par_threads(0); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2104 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2105 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2106 class G1RefEnqueueTaskProxy: public AbstractGangTask { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2107 typedef AbstractRefProcTaskExecutor::EnqueueTask EnqueueTask; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2108 EnqueueTask& _enq_task; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2109 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2110 public: |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2111 G1RefEnqueueTaskProxy(EnqueueTask& enq_task) : |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2112 AbstractGangTask("Enqueue reference objects in parallel"), |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2113 _enq_task(enq_task) |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2114 { } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2115 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2116 virtual void work(int i) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2117 _enq_task.work(i); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2118 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2119 }; |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2120 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2121 void G1RefProcTaskExecutor::execute(EnqueueTask& enq_task) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2122 assert(_workers != NULL, "Need parallel worker threads."); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2123 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2124 G1RefEnqueueTaskProxy enq_task_proxy(enq_task); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2125 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2126 _g1h->set_par_threads(_active_workers); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2127 _workers->run_task(&enq_task_proxy); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2128 _g1h->set_par_threads(0); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2129 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2130 |
342 | 2131 void ConcurrentMark::weakRefsWork(bool clear_all_soft_refs) { |
2132 ResourceMark rm; | |
2133 HandleMark hm; | |
453
c96030fff130
6684579: SoftReference processing can be made more efficient
ysr
parents:
355
diff
changeset
|
2134 G1CollectedHeap* g1h = G1CollectedHeap::heap(); |
c96030fff130
6684579: SoftReference processing can be made more efficient
ysr
parents:
355
diff
changeset
|
2135 ReferenceProcessor* rp = g1h->ref_processor(); |
342 | 2136 |
1974
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
2137 // See the comment in G1CollectedHeap::ref_processing_init() |
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
2138 // about how reference processing currently works in G1. |
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
2139 |
342 | 2140 // Process weak references. |
457
27a80744a83b
6778647: snap(), snap_policy() should be renamed setup(), setup_policy()
ysr
parents:
453
diff
changeset
|
2141 rp->setup_policy(clear_all_soft_refs); |
342 | 2142 assert(_markStack.isEmpty(), "mark stack should be empty"); |
2143 | |
2037
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
2144 G1CMIsAliveClosure g1_is_alive(g1h); |
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
2145 G1CMKeepAliveClosure g1_keep_alive(g1h, this, nextMarkBitMap()); |
342 | 2146 G1CMDrainMarkingStackClosure |
2037
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
2147 g1_drain_mark_stack(nextMarkBitMap(), &_markStack, &g1_keep_alive); |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2148 // We use the work gang from the G1CollectedHeap and we utilize all |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2149 // the worker threads. |
2369
92da084fefc9
6668573: CMS: reference processing crash if ParallelCMSThreads > ParallelGCThreads
ysr
parents:
2361
diff
changeset
|
2150 int active_workers = g1h->workers() ? g1h->workers()->total_workers() : 1; |
92da084fefc9
6668573: CMS: reference processing crash if ParallelCMSThreads > ParallelGCThreads
ysr
parents:
2361
diff
changeset
|
2151 active_workers = MAX2(MIN2(active_workers, (int)_max_task_num), 1); |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2152 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2153 G1RefProcTaskExecutor par_task_executor(g1h, this, nextMarkBitMap(), |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2154 g1h->workers(), active_workers); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2155 |
2369
92da084fefc9
6668573: CMS: reference processing crash if ParallelCMSThreads > ParallelGCThreads
ysr
parents:
2361
diff
changeset
|
2156 |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2157 if (rp->processing_is_mt()) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2158 // Set the degree of MT here. If the discovery is done MT, there |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2159 // may have been a different number of threads doing the discovery |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2160 // and a different number of discovered lists may have Ref objects. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2161 // That is OK as long as the Reference lists are balanced (see |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2162 // balance_all_queues() and balance_queues()). |
2369
92da084fefc9
6668573: CMS: reference processing crash if ParallelCMSThreads > ParallelGCThreads
ysr
parents:
2361
diff
changeset
|
2163 rp->set_active_mt_degree(active_workers); |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2164 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2165 rp->process_discovered_references(&g1_is_alive, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2166 &g1_keep_alive, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2167 &g1_drain_mark_stack, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2168 &par_task_executor); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2169 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2170 // The work routines of the parallel keep_alive and drain_marking_stack |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2171 // will set the has_overflown flag if we overflow the global marking |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2172 // stack. |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2173 } else { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2174 rp->process_discovered_references(&g1_is_alive, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2175 &g1_keep_alive, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2176 &g1_drain_mark_stack, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2177 NULL); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2178 |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2179 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2180 |
342 | 2181 assert(_markStack.overflow() || _markStack.isEmpty(), |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2182 "mark stack should be empty (unless it overflowed)"); |
342 | 2183 if (_markStack.overflow()) { |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2184 // Should have been done already when we tried to push an |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2185 // entry on to the global mark stack. But let's do it again. |
342 | 2186 set_has_overflown(); |
2187 } | |
2188 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2189 if (rp->processing_is_mt()) { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2190 assert(rp->num_q() == active_workers, "why not"); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2191 rp->enqueue_discovered_references(&par_task_executor); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2192 } else { |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2193 rp->enqueue_discovered_references(); |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2194 } |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2195 |
342 | 2196 rp->verify_no_references_recorded(); |
2197 assert(!rp->discovery_enabled(), "should have been disabled"); | |
2198 | |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
2175
diff
changeset
|
2199 // Now clean up stale oops in StringTable |
2037
b03260081e9b
7006113: G1: Initialize ReferenceProcessor::_is_alive_non_header field
johnc
parents:
1974
diff
changeset
|
2200 StringTable::unlink(&g1_is_alive); |
2177
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
2175
diff
changeset
|
2201 // Clean up unreferenced symbols in symbol table. |
3582bf76420e
6990754: Use native memory and reference counting to implement SymbolTable
coleenp
parents:
2175
diff
changeset
|
2202 SymbolTable::unlink(); |
342 | 2203 } |
2204 | |
2205 void ConcurrentMark::swapMarkBitMaps() { | |
2206 CMBitMapRO* temp = _prevMarkBitMap; | |
2207 _prevMarkBitMap = (CMBitMapRO*)_nextMarkBitMap; | |
2208 _nextMarkBitMap = (CMBitMap*) temp; | |
2209 } | |
2210 | |
2211 class CMRemarkTask: public AbstractGangTask { | |
2212 private: | |
2213 ConcurrentMark *_cm; | |
2214 | |
2215 public: | |
2216 void work(int worker_i) { | |
2217 // Since all available tasks are actually started, we should | |
2218 // only proceed if we're supposed to be actived. | |
2219 if ((size_t)worker_i < _cm->active_tasks()) { | |
2220 CMTask* task = _cm->task(worker_i); | |
2221 task->record_start_time(); | |
2222 do { | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2223 task->do_marking_step(1000000000.0 /* something very large */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2224 true /* do_stealing */, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2225 true /* do_termination */); |
342 | 2226 } while (task->has_aborted() && !_cm->has_overflown()); |
2227 // If we overflow, then we do not want to restart. We instead | |
2228 // want to abort remark and do concurrent marking again. | |
2229 task->record_end_time(); | |
2230 } | |
2231 } | |
2232 | |
2233 CMRemarkTask(ConcurrentMark* cm) : | |
2234 AbstractGangTask("Par Remark"), _cm(cm) { } | |
2235 }; | |
2236 | |
2237 void ConcurrentMark::checkpointRootsFinalWork() { | |
2238 ResourceMark rm; | |
2239 HandleMark hm; | |
2240 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
2241 | |
2242 g1h->ensure_parsability(false); | |
2243 | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
2244 if (G1CollectedHeap::use_parallel_gc_threads()) { |
989
148e5441d916
6863023: need non-perm oops in code cache for JSR 292
jrose
parents:
912
diff
changeset
|
2245 G1CollectedHeap::StrongRootsScope srs(g1h); |
342 | 2246 // this is remark, so we'll use up all available threads |
2247 int active_workers = ParallelGCThreads; | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2248 set_phase(active_workers, false /* concurrent */); |
342 | 2249 |
2250 CMRemarkTask remarkTask(this); | |
2251 // We will start all available threads, even if we decide that the | |
2252 // active_workers will be fewer. The extra ones will just bail out | |
2253 // immediately. | |
2254 int n_workers = g1h->workers()->total_workers(); | |
2255 g1h->set_par_threads(n_workers); | |
2256 g1h->workers()->run_task(&remarkTask); | |
2257 g1h->set_par_threads(0); | |
2258 } else { | |
989
148e5441d916
6863023: need non-perm oops in code cache for JSR 292
jrose
parents:
912
diff
changeset
|
2259 G1CollectedHeap::StrongRootsScope srs(g1h); |
342 | 2260 // this is remark, so we'll use up all available threads |
2261 int active_workers = 1; | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
2262 set_phase(active_workers, false /* concurrent */); |
342 | 2263 |
2264 CMRemarkTask remarkTask(this); | |
2265 // We will start all available threads, even if we decide that the | |
2266 // active_workers will be fewer. The extra ones will just bail out | |
2267 // immediately. | |
2268 remarkTask.work(0); | |
2269 } | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2270 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2271 guarantee(satb_mq_set.completed_buffers_num() == 0, "invariant"); |
342 | 2272 |
2273 print_stats(); | |
2274 | |
2275 #if VERIFY_OBJS_PROCESSED | |
2276 if (_scan_obj_cl.objs_processed != ThreadLocalObjQueue::objs_enqueued) { | |
2277 gclog_or_tty->print_cr("Processed = %d, enqueued = %d.", | |
2278 _scan_obj_cl.objs_processed, | |
2279 ThreadLocalObjQueue::objs_enqueued); | |
2280 guarantee(_scan_obj_cl.objs_processed == | |
2281 ThreadLocalObjQueue::objs_enqueued, | |
2282 "Different number of objs processed and enqueued."); | |
2283 } | |
2284 #endif | |
2285 } | |
2286 | |
1044 | 2287 #ifndef PRODUCT |
2288 | |
1388 | 2289 class PrintReachableOopClosure: public OopClosure { |
342 | 2290 private: |
2291 G1CollectedHeap* _g1h; | |
2292 CMBitMapRO* _bitmap; | |
2293 outputStream* _out; | |
1044 | 2294 bool _use_prev_marking; |
1388 | 2295 bool _all; |
342 | 2296 |
2297 public: | |
1388 | 2298 PrintReachableOopClosure(CMBitMapRO* bitmap, |
2299 outputStream* out, | |
2300 bool use_prev_marking, | |
2301 bool all) : | |
1044 | 2302 _g1h(G1CollectedHeap::heap()), |
1388 | 2303 _bitmap(bitmap), _out(out), _use_prev_marking(use_prev_marking), _all(all) { } |
342 | 2304 |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2305 void do_oop(narrowOop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2306 void do_oop( oop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2307 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2308 template <class T> void do_oop_work(T* p) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2309 oop obj = oopDesc::load_decode_heap_oop(p); |
342 | 2310 const char* str = NULL; |
2311 const char* str2 = ""; | |
2312 | |
1388 | 2313 if (obj == NULL) { |
2314 str = ""; | |
2315 } else if (!_g1h->is_in_g1_reserved(obj)) { | |
2316 str = " O"; | |
2317 } else { | |
342 | 2318 HeapRegion* hr = _g1h->heap_region_containing(obj); |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2319 guarantee(hr != NULL, "invariant"); |
1044 | 2320 bool over_tams = false; |
2321 if (_use_prev_marking) { | |
2322 over_tams = hr->obj_allocated_since_prev_marking(obj); | |
2323 } else { | |
2324 over_tams = hr->obj_allocated_since_next_marking(obj); | |
2325 } | |
1388 | 2326 bool marked = _bitmap->isMarked((HeapWord*) obj); |
1044 | 2327 |
2328 if (over_tams) { | |
1388 | 2329 str = " >"; |
2330 if (marked) { | |
342 | 2331 str2 = " AND MARKED"; |
1044 | 2332 } |
1388 | 2333 } else if (marked) { |
2334 str = " M"; | |
1044 | 2335 } else { |
1388 | 2336 str = " NOT"; |
1044 | 2337 } |
342 | 2338 } |
2339 | |
1388 | 2340 _out->print_cr(" "PTR_FORMAT": "PTR_FORMAT"%s%s", |
342 | 2341 p, (void*) obj, str, str2); |
2342 } | |
2343 }; | |
2344 | |
1388 | 2345 class PrintReachableObjectClosure : public ObjectClosure { |
342 | 2346 private: |
1044 | 2347 CMBitMapRO* _bitmap; |
342 | 2348 outputStream* _out; |
1044 | 2349 bool _use_prev_marking; |
1388 | 2350 bool _all; |
2351 HeapRegion* _hr; | |
342 | 2352 |
2353 public: | |
1388 | 2354 PrintReachableObjectClosure(CMBitMapRO* bitmap, |
2355 outputStream* out, | |
2356 bool use_prev_marking, | |
2357 bool all, | |
2358 HeapRegion* hr) : | |
2359 _bitmap(bitmap), _out(out), | |
2360 _use_prev_marking(use_prev_marking), _all(all), _hr(hr) { } | |
2361 | |
2362 void do_object(oop o) { | |
2363 bool over_tams; | |
2364 if (_use_prev_marking) { | |
2365 over_tams = _hr->obj_allocated_since_prev_marking(o); | |
2366 } else { | |
2367 over_tams = _hr->obj_allocated_since_next_marking(o); | |
2368 } | |
2369 bool marked = _bitmap->isMarked((HeapWord*) o); | |
2370 bool print_it = _all || over_tams || marked; | |
2371 | |
2372 if (print_it) { | |
2373 _out->print_cr(" "PTR_FORMAT"%s", | |
2374 o, (over_tams) ? " >" : (marked) ? " M" : ""); | |
2375 PrintReachableOopClosure oopCl(_bitmap, _out, _use_prev_marking, _all); | |
2376 o->oop_iterate(&oopCl); | |
2377 } | |
342 | 2378 } |
2379 }; | |
2380 | |
1388 | 2381 class PrintReachableRegionClosure : public HeapRegionClosure { |
342 | 2382 private: |
1044 | 2383 CMBitMapRO* _bitmap; |
342 | 2384 outputStream* _out; |
1044 | 2385 bool _use_prev_marking; |
1388 | 2386 bool _all; |
342 | 2387 |
2388 public: | |
2389 bool doHeapRegion(HeapRegion* hr) { | |
2390 HeapWord* b = hr->bottom(); | |
2391 HeapWord* e = hr->end(); | |
2392 HeapWord* t = hr->top(); | |
1044 | 2393 HeapWord* p = NULL; |
2394 if (_use_prev_marking) { | |
2395 p = hr->prev_top_at_mark_start(); | |
2396 } else { | |
2397 p = hr->next_top_at_mark_start(); | |
2398 } | |
342 | 2399 _out->print_cr("** ["PTR_FORMAT", "PTR_FORMAT"] top: "PTR_FORMAT" " |
1044 | 2400 "TAMS: "PTR_FORMAT, b, e, t, p); |
1388 | 2401 _out->cr(); |
2402 | |
2403 HeapWord* from = b; | |
2404 HeapWord* to = t; | |
2405 | |
2406 if (to > from) { | |
2407 _out->print_cr("Objects in ["PTR_FORMAT", "PTR_FORMAT"]", from, to); | |
2408 _out->cr(); | |
2409 PrintReachableObjectClosure ocl(_bitmap, _out, | |
2410 _use_prev_marking, _all, hr); | |
2411 hr->object_iterate_mem_careful(MemRegion(from, to), &ocl); | |
2412 _out->cr(); | |
2413 } | |
342 | 2414 |
2415 return false; | |
2416 } | |
2417 | |
1388 | 2418 PrintReachableRegionClosure(CMBitMapRO* bitmap, |
2419 outputStream* out, | |
2420 bool use_prev_marking, | |
2421 bool all) : | |
2422 _bitmap(bitmap), _out(out), _use_prev_marking(use_prev_marking), _all(all) { } | |
342 | 2423 }; |
2424 | |
1388 | 2425 void ConcurrentMark::print_reachable(const char* str, |
2426 bool use_prev_marking, | |
2427 bool all) { | |
2428 gclog_or_tty->cr(); | |
2429 gclog_or_tty->print_cr("== Doing heap dump... "); | |
1044 | 2430 |
2431 if (G1PrintReachableBaseFile == NULL) { | |
2432 gclog_or_tty->print_cr(" #### error: no base file defined"); | |
2433 return; | |
2434 } | |
2435 | |
2436 if (strlen(G1PrintReachableBaseFile) + 1 + strlen(str) > | |
2437 (JVM_MAXPATHLEN - 1)) { | |
2438 gclog_or_tty->print_cr(" #### error: file name too long"); | |
2439 return; | |
2440 } | |
2441 | |
2442 char file_name[JVM_MAXPATHLEN]; | |
2443 sprintf(file_name, "%s.%s", G1PrintReachableBaseFile, str); | |
2444 gclog_or_tty->print_cr(" dumping to file %s", file_name); | |
2445 | |
2446 fileStream fout(file_name); | |
2447 if (!fout.is_open()) { | |
2448 gclog_or_tty->print_cr(" #### error: could not open file"); | |
2449 return; | |
2450 } | |
2451 | |
2452 outputStream* out = &fout; | |
2453 | |
2454 CMBitMapRO* bitmap = NULL; | |
2455 if (use_prev_marking) { | |
2456 bitmap = _prevMarkBitMap; | |
2457 } else { | |
2458 bitmap = _nextMarkBitMap; | |
2459 } | |
2460 | |
2461 out->print_cr("-- USING %s", (use_prev_marking) ? "PTAMS" : "NTAMS"); | |
2462 out->cr(); | |
2463 | |
1388 | 2464 out->print_cr("--- ITERATING OVER REGIONS"); |
1044 | 2465 out->cr(); |
1388 | 2466 PrintReachableRegionClosure rcl(bitmap, out, use_prev_marking, all); |
2467 _g1h->heap_region_iterate(&rcl); | |
1044 | 2468 out->cr(); |
2469 | |
2470 gclog_or_tty->print_cr(" done"); | |
1388 | 2471 gclog_or_tty->flush(); |
342 | 2472 } |
2473 | |
1044 | 2474 #endif // PRODUCT |
2475 | |
342 | 2476 // This note is for drainAllSATBBuffers and the code in between. |
2477 // In the future we could reuse a task to do this work during an | |
2478 // evacuation pause (since now tasks are not active and can be claimed | |
2479 // during an evacuation pause). This was a late change to the code and | |
2480 // is currently not being taken advantage of. | |
2481 | |
2482 class CMGlobalObjectClosure : public ObjectClosure { | |
2483 private: | |
2484 ConcurrentMark* _cm; | |
2485 | |
2486 public: | |
2487 void do_object(oop obj) { | |
2488 _cm->deal_with_reference(obj); | |
2489 } | |
2490 | |
2491 CMGlobalObjectClosure(ConcurrentMark* cm) : _cm(cm) { } | |
2492 }; | |
2493 | |
2494 void ConcurrentMark::deal_with_reference(oop obj) { | |
2495 if (verbose_high()) | |
2496 gclog_or_tty->print_cr("[global] we're dealing with reference "PTR_FORMAT, | |
2497 (void*) obj); | |
2498 | |
2499 | |
2500 HeapWord* objAddr = (HeapWord*) obj; | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2501 assert(obj->is_oop_or_null(true /* ignore mark word */), "Error"); |
342 | 2502 if (_g1h->is_in_g1_reserved(objAddr)) { |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2503 assert(obj != NULL, "is_in_g1_reserved should ensure this"); |
342 | 2504 HeapRegion* hr = _g1h->heap_region_containing(obj); |
2505 if (_g1h->is_obj_ill(obj, hr)) { | |
2506 if (verbose_high()) | |
2507 gclog_or_tty->print_cr("[global] "PTR_FORMAT" is not considered " | |
2508 "marked", (void*) obj); | |
2509 | |
2510 // we need to mark it first | |
2511 if (_nextMarkBitMap->parMark(objAddr)) { | |
2512 // No OrderAccess:store_load() is needed. It is implicit in the | |
2513 // CAS done in parMark(objAddr) above | |
2514 HeapWord* finger = _finger; | |
2515 if (objAddr < finger) { | |
2516 if (verbose_high()) | |
2517 gclog_or_tty->print_cr("[global] below the global finger " | |
2518 "("PTR_FORMAT"), pushing it", finger); | |
2519 if (!mark_stack_push(obj)) { | |
2520 if (verbose_low()) | |
2521 gclog_or_tty->print_cr("[global] global stack overflow during " | |
2522 "deal_with_reference"); | |
2523 } | |
2524 } | |
2525 } | |
2526 } | |
2527 } | |
2528 } | |
2529 | |
2530 void ConcurrentMark::drainAllSATBBuffers() { | |
2531 CMGlobalObjectClosure oc(this); | |
2532 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); | |
2533 satb_mq_set.set_closure(&oc); | |
2534 | |
2535 while (satb_mq_set.apply_closure_to_completed_buffer()) { | |
2536 if (verbose_medium()) | |
2537 gclog_or_tty->print_cr("[global] processed an SATB buffer"); | |
2538 } | |
2539 | |
2540 // no need to check whether we should do this, as this is only | |
2541 // called during an evacuation pause | |
2542 satb_mq_set.iterate_closure_all_threads(); | |
2543 | |
2544 satb_mq_set.set_closure(NULL); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2545 assert(satb_mq_set.completed_buffers_num() == 0, "invariant"); |
342 | 2546 } |
2547 | |
2548 void ConcurrentMark::markPrev(oop p) { | |
2549 // Note we are overriding the read-only view of the prev map here, via | |
2550 // the cast. | |
2551 ((CMBitMap*)_prevMarkBitMap)->mark((HeapWord*)p); | |
2552 } | |
2553 | |
2554 void ConcurrentMark::clear(oop p) { | |
2555 assert(p != NULL && p->is_oop(), "expected an oop"); | |
2556 HeapWord* addr = (HeapWord*)p; | |
2557 assert(addr >= _nextMarkBitMap->startWord() || | |
2558 addr < _nextMarkBitMap->endWord(), "in a region"); | |
2559 | |
2560 _nextMarkBitMap->clear(addr); | |
2561 } | |
2562 | |
2563 void ConcurrentMark::clearRangeBothMaps(MemRegion mr) { | |
2564 // Note we are overriding the read-only view of the prev map here, via | |
2565 // the cast. | |
2566 ((CMBitMap*)_prevMarkBitMap)->clearRange(mr); | |
2567 _nextMarkBitMap->clearRange(mr); | |
2568 } | |
2569 | |
2570 HeapRegion* | |
2571 ConcurrentMark::claim_region(int task_num) { | |
2572 // "checkpoint" the finger | |
2573 HeapWord* finger = _finger; | |
2574 | |
2575 // _heap_end will not change underneath our feet; it only changes at | |
2576 // yield points. | |
2577 while (finger < _heap_end) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2578 assert(_g1h->is_in_g1_reserved(finger), "invariant"); |
342 | 2579 |
2580 // is the gap between reading the finger and doing the CAS too long? | |
2581 | |
2582 HeapRegion* curr_region = _g1h->heap_region_containing(finger); | |
2583 HeapWord* bottom = curr_region->bottom(); | |
2584 HeapWord* end = curr_region->end(); | |
2585 HeapWord* limit = curr_region->next_top_at_mark_start(); | |
2586 | |
2587 if (verbose_low()) | |
2588 gclog_or_tty->print_cr("[%d] curr_region = "PTR_FORMAT" " | |
2589 "["PTR_FORMAT", "PTR_FORMAT"), " | |
2590 "limit = "PTR_FORMAT, | |
2591 task_num, curr_region, bottom, end, limit); | |
2592 | |
2593 HeapWord* res = | |
2594 (HeapWord*) Atomic::cmpxchg_ptr(end, &_finger, finger); | |
2595 if (res == finger) { | |
2596 // we succeeded | |
2597 | |
2598 // notice that _finger == end cannot be guaranteed here since, | |
2599 // someone else might have moved the finger even further | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2600 assert(_finger >= end, "the finger should have moved forward"); |
342 | 2601 |
2602 if (verbose_low()) | |
2603 gclog_or_tty->print_cr("[%d] we were successful with region = " | |
2604 PTR_FORMAT, task_num, curr_region); | |
2605 | |
2606 if (limit > bottom) { | |
2607 if (verbose_low()) | |
2608 gclog_or_tty->print_cr("[%d] region "PTR_FORMAT" is not empty, " | |
2609 "returning it ", task_num, curr_region); | |
2610 return curr_region; | |
2611 } else { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2612 assert(limit == bottom, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2613 "the region limit should be at bottom"); |
342 | 2614 if (verbose_low()) |
2615 gclog_or_tty->print_cr("[%d] region "PTR_FORMAT" is empty, " | |
2616 "returning NULL", task_num, curr_region); | |
2617 // we return NULL and the caller should try calling | |
2618 // claim_region() again. | |
2619 return NULL; | |
2620 } | |
2621 } else { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2622 assert(_finger > finger, "the finger should have moved forward"); |
342 | 2623 if (verbose_low()) |
2624 gclog_or_tty->print_cr("[%d] somebody else moved the finger, " | |
2625 "global finger = "PTR_FORMAT", " | |
2626 "our finger = "PTR_FORMAT, | |
2627 task_num, _finger, finger); | |
2628 | |
2629 // read it again | |
2630 finger = _finger; | |
2631 } | |
2632 } | |
2633 | |
2634 return NULL; | |
2635 } | |
2636 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2637 bool ConcurrentMark::invalidate_aborted_regions_in_cset() { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2638 bool result = false; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2639 for (int i = 0; i < (int)_max_task_num; ++i) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2640 CMTask* the_task = _tasks[i]; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2641 MemRegion mr = the_task->aborted_region(); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2642 if (mr.start() != NULL) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2643 assert(mr.end() != NULL, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2644 assert(mr.word_size() > 0, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2645 HeapRegion* hr = _g1h->heap_region_containing(mr.start()); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2646 assert(hr != NULL, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2647 if (hr->in_collection_set()) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2648 // The region points into the collection set |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2649 the_task->set_aborted_region(MemRegion()); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2650 result = true; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2651 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2652 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2653 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2654 return result; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2655 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2656 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2657 bool ConcurrentMark::has_aborted_regions() { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2658 for (int i = 0; i < (int)_max_task_num; ++i) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2659 CMTask* the_task = _tasks[i]; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2660 MemRegion mr = the_task->aborted_region(); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2661 if (mr.start() != NULL) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2662 assert(mr.end() != NULL, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2663 assert(mr.word_size() > 0, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2664 return true; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2665 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2666 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2667 return false; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2668 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2669 |
342 | 2670 void ConcurrentMark::oops_do(OopClosure* cl) { |
2671 if (_markStack.size() > 0 && verbose_low()) | |
2672 gclog_or_tty->print_cr("[global] scanning the global marking stack, " | |
2673 "size = %d", _markStack.size()); | |
2674 // we first iterate over the contents of the mark stack... | |
2675 _markStack.oops_do(cl); | |
2676 | |
2677 for (int i = 0; i < (int)_max_task_num; ++i) { | |
2678 OopTaskQueue* queue = _task_queues->queue((int)i); | |
2679 | |
2680 if (queue->size() > 0 && verbose_low()) | |
2681 gclog_or_tty->print_cr("[global] scanning task queue of task %d, " | |
2682 "size = %d", i, queue->size()); | |
2683 | |
2684 // ...then over the contents of the all the task queues. | |
2685 queue->oops_do(cl); | |
2686 } | |
2687 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2688 // Invalidate any entries, that are in the region stack, that |
342 | 2689 // point into the collection set |
2690 if (_regionStack.invalidate_entries_into_cset()) { | |
2691 // otherwise, any gray objects copied during the evacuation pause | |
2692 // might not be visited. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
2693 assert(_should_gray_objects, "invariant"); |
342 | 2694 } |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2695 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2696 // Invalidate any aborted regions, recorded in the individual CM |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2697 // tasks, that point into the collection set. |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2698 if (invalidate_aborted_regions_in_cset()) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2699 // otherwise, any gray objects copied during the evacuation pause |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2700 // might not be visited. |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2701 assert(_should_gray_objects, "invariant"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2702 } |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2703 |
342 | 2704 } |
2705 | |
2706 void ConcurrentMark::clear_marking_state() { | |
2707 _markStack.setEmpty(); | |
2708 _markStack.clear_overflow(); | |
2709 _regionStack.setEmpty(); | |
2710 _regionStack.clear_overflow(); | |
2711 clear_has_overflown(); | |
2712 _finger = _heap_start; | |
2713 | |
2714 for (int i = 0; i < (int)_max_task_num; ++i) { | |
2715 OopTaskQueue* queue = _task_queues->queue(i); | |
2716 queue->set_empty(); | |
1885
a5c514e74487
6988458: G1: assert(mr.end() <= _cm->finger()) failed: otherwise the region shouldn't be on the stack
johnc
parents:
1835
diff
changeset
|
2717 // Clear any partial regions from the CMTasks |
a5c514e74487
6988458: G1: assert(mr.end() <= _cm->finger()) failed: otherwise the region shouldn't be on the stack
johnc
parents:
1835
diff
changeset
|
2718 _tasks[i]->clear_aborted_region(); |
342 | 2719 } |
2720 } | |
2721 | |
2722 void ConcurrentMark::print_stats() { | |
2723 if (verbose_stats()) { | |
2724 gclog_or_tty->print_cr("---------------------------------------------------------------------"); | |
2725 for (size_t i = 0; i < _active_tasks; ++i) { | |
2726 _tasks[i]->print_stats(); | |
2727 gclog_or_tty->print_cr("---------------------------------------------------------------------"); | |
2728 } | |
2729 } | |
2730 } | |
2731 | |
2732 class CSMarkOopClosure: public OopClosure { | |
2733 friend class CSMarkBitMapClosure; | |
2734 | |
2735 G1CollectedHeap* _g1h; | |
2736 CMBitMap* _bm; | |
2737 ConcurrentMark* _cm; | |
2738 oop* _ms; | |
2739 jint* _array_ind_stack; | |
2740 int _ms_size; | |
2741 int _ms_ind; | |
2742 int _array_increment; | |
2743 | |
2744 bool push(oop obj, int arr_ind = 0) { | |
2745 if (_ms_ind == _ms_size) { | |
2746 gclog_or_tty->print_cr("Mark stack is full."); | |
2747 return false; | |
2748 } | |
2749 _ms[_ms_ind] = obj; | |
2750 if (obj->is_objArray()) _array_ind_stack[_ms_ind] = arr_ind; | |
2751 _ms_ind++; | |
2752 return true; | |
2753 } | |
2754 | |
2755 oop pop() { | |
2756 if (_ms_ind == 0) return NULL; | |
2757 else { | |
2758 _ms_ind--; | |
2759 return _ms[_ms_ind]; | |
2760 } | |
2761 } | |
2762 | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2763 template <class T> bool drain() { |
342 | 2764 while (_ms_ind > 0) { |
2765 oop obj = pop(); | |
2766 assert(obj != NULL, "Since index was non-zero."); | |
2767 if (obj->is_objArray()) { | |
2768 jint arr_ind = _array_ind_stack[_ms_ind]; | |
2769 objArrayOop aobj = objArrayOop(obj); | |
2770 jint len = aobj->length(); | |
2771 jint next_arr_ind = arr_ind + _array_increment; | |
2772 if (next_arr_ind < len) { | |
2773 push(obj, next_arr_ind); | |
2774 } | |
2775 // Now process this portion of this one. | |
2776 int lim = MIN2(next_arr_ind, len); | |
2777 for (int j = arr_ind; j < lim; j++) { | |
912
308762b2bf14
6872000: G1: compilation fails on linux/older gcc
apetrusenko
parents:
866
diff
changeset
|
2778 do_oop(aobj->objArrayOopDesc::obj_at_addr<T>(j)); |
342 | 2779 } |
2780 | |
2781 } else { | |
2782 obj->oop_iterate(this); | |
2783 } | |
2784 if (abort()) return false; | |
2785 } | |
2786 return true; | |
2787 } | |
2788 | |
2789 public: | |
2790 CSMarkOopClosure(ConcurrentMark* cm, int ms_size) : | |
2791 _g1h(G1CollectedHeap::heap()), | |
2792 _cm(cm), | |
2793 _bm(cm->nextMarkBitMap()), | |
2794 _ms_size(ms_size), _ms_ind(0), | |
2795 _ms(NEW_C_HEAP_ARRAY(oop, ms_size)), | |
2796 _array_ind_stack(NEW_C_HEAP_ARRAY(jint, ms_size)), | |
2797 _array_increment(MAX2(ms_size/8, 16)) | |
2798 {} | |
2799 | |
2800 ~CSMarkOopClosure() { | |
2801 FREE_C_HEAP_ARRAY(oop, _ms); | |
2802 FREE_C_HEAP_ARRAY(jint, _array_ind_stack); | |
2803 } | |
2804 | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2805 virtual void do_oop(narrowOop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2806 virtual void do_oop( oop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2807 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2808 template <class T> void do_oop_work(T* p) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2809 T heap_oop = oopDesc::load_heap_oop(p); |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2810 if (oopDesc::is_null(heap_oop)) return; |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2811 oop obj = oopDesc::decode_heap_oop_not_null(heap_oop); |
342 | 2812 if (obj->is_forwarded()) { |
2813 // If the object has already been forwarded, we have to make sure | |
2814 // that it's marked. So follow the forwarding pointer. Note that | |
2815 // this does the right thing for self-forwarding pointers in the | |
2816 // evacuation failure case. | |
2817 obj = obj->forwardee(); | |
2818 } | |
2819 HeapRegion* hr = _g1h->heap_region_containing(obj); | |
2820 if (hr != NULL) { | |
2821 if (hr->in_collection_set()) { | |
2822 if (_g1h->is_obj_ill(obj)) { | |
2823 _bm->mark((HeapWord*)obj); | |
2824 if (!push(obj)) { | |
2825 gclog_or_tty->print_cr("Setting abort in CSMarkOopClosure because push failed."); | |
2826 set_abort(); | |
2827 } | |
2828 } | |
2829 } else { | |
2830 // Outside the collection set; we need to gray it | |
2831 _cm->deal_with_reference(obj); | |
2832 } | |
2833 } | |
2834 } | |
2835 }; | |
2836 | |
2837 class CSMarkBitMapClosure: public BitMapClosure { | |
2838 G1CollectedHeap* _g1h; | |
2839 CMBitMap* _bitMap; | |
2840 ConcurrentMark* _cm; | |
2841 CSMarkOopClosure _oop_cl; | |
2842 public: | |
2843 CSMarkBitMapClosure(ConcurrentMark* cm, int ms_size) : | |
2844 _g1h(G1CollectedHeap::heap()), | |
2845 _bitMap(cm->nextMarkBitMap()), | |
2846 _oop_cl(cm, ms_size) | |
2847 {} | |
2848 | |
2849 ~CSMarkBitMapClosure() {} | |
2850 | |
2851 bool do_bit(size_t offset) { | |
2852 // convert offset into a HeapWord* | |
2853 HeapWord* addr = _bitMap->offsetToHeapWord(offset); | |
2854 assert(_bitMap->endWord() && addr < _bitMap->endWord(), | |
2855 "address out of range"); | |
2856 assert(_bitMap->isMarked(addr), "tautology"); | |
2857 oop obj = oop(addr); | |
2858 if (!obj->is_forwarded()) { | |
2859 if (!_oop_cl.push(obj)) return false; | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2860 if (UseCompressedOops) { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2861 if (!_oop_cl.drain<narrowOop>()) return false; |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2862 } else { |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2863 if (!_oop_cl.drain<oop>()) return false; |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
2864 } |
342 | 2865 } |
2866 // Otherwise... | |
2867 return true; | |
2868 } | |
2869 }; | |
2870 | |
2871 | |
2872 class CompleteMarkingInCSHRClosure: public HeapRegionClosure { | |
2873 CMBitMap* _bm; | |
2874 CSMarkBitMapClosure _bit_cl; | |
2875 enum SomePrivateConstants { | |
2876 MSSize = 1000 | |
2877 }; | |
2878 bool _completed; | |
2879 public: | |
2880 CompleteMarkingInCSHRClosure(ConcurrentMark* cm) : | |
2881 _bm(cm->nextMarkBitMap()), | |
2882 _bit_cl(cm, MSSize), | |
2883 _completed(true) | |
2884 {} | |
2885 | |
2886 ~CompleteMarkingInCSHRClosure() {} | |
2887 | |
2888 bool doHeapRegion(HeapRegion* r) { | |
2889 if (!r->evacuation_failed()) { | |
2890 MemRegion mr = MemRegion(r->bottom(), r->next_top_at_mark_start()); | |
2891 if (!mr.is_empty()) { | |
2892 if (!_bm->iterate(&_bit_cl, mr)) { | |
2893 _completed = false; | |
2894 return true; | |
2895 } | |
2896 } | |
2897 } | |
2898 return false; | |
2899 } | |
2900 | |
2901 bool completed() { return _completed; } | |
2902 }; | |
2903 | |
2904 class ClearMarksInHRClosure: public HeapRegionClosure { | |
2905 CMBitMap* _bm; | |
2906 public: | |
2907 ClearMarksInHRClosure(CMBitMap* bm): _bm(bm) { } | |
2908 | |
2909 bool doHeapRegion(HeapRegion* r) { | |
2910 if (!r->used_region().is_empty() && !r->evacuation_failed()) { | |
2911 MemRegion usedMR = r->used_region(); | |
2912 _bm->clearRange(r->used_region()); | |
2913 } | |
2914 return false; | |
2915 } | |
2916 }; | |
2917 | |
2918 void ConcurrentMark::complete_marking_in_collection_set() { | |
2919 G1CollectedHeap* g1h = G1CollectedHeap::heap(); | |
2920 | |
2921 if (!g1h->mark_in_progress()) { | |
2922 g1h->g1_policy()->record_mark_closure_time(0.0); | |
2923 return; | |
2924 } | |
2925 | |
2926 int i = 1; | |
2927 double start = os::elapsedTime(); | |
2928 while (true) { | |
2929 i++; | |
2930 CompleteMarkingInCSHRClosure cmplt(this); | |
2931 g1h->collection_set_iterate(&cmplt); | |
2932 if (cmplt.completed()) break; | |
2933 } | |
2934 double end_time = os::elapsedTime(); | |
2935 double elapsed_time_ms = (end_time - start) * 1000.0; | |
2936 g1h->g1_policy()->record_mark_closure_time(elapsed_time_ms); | |
2937 | |
2938 ClearMarksInHRClosure clr(nextMarkBitMap()); | |
2939 g1h->collection_set_iterate(&clr); | |
2940 } | |
2941 | |
2942 // The next two methods deal with the following optimisation. Some | |
2943 // objects are gray by being marked and located above the finger. If | |
2944 // they are copied, during an evacuation pause, below the finger then | |
2945 // the need to be pushed on the stack. The observation is that, if | |
2946 // there are no regions in the collection set located above the | |
2947 // finger, then the above cannot happen, hence we do not need to | |
2948 // explicitly gray any objects when copying them to below the | |
2949 // finger. The global stack will be scanned to ensure that, if it | |
2950 // points to objects being copied, it will update their | |
2951 // location. There is a tricky situation with the gray objects in | |
2952 // region stack that are being coped, however. See the comment in | |
2953 // newCSet(). | |
2954 | |
2955 void ConcurrentMark::newCSet() { | |
2956 if (!concurrent_marking_in_progress()) | |
2957 // nothing to do if marking is not in progress | |
2958 return; | |
2959 | |
2960 // find what the lowest finger is among the global and local fingers | |
2961 _min_finger = _finger; | |
2962 for (int i = 0; i < (int)_max_task_num; ++i) { | |
2963 CMTask* task = _tasks[i]; | |
2964 HeapWord* task_finger = task->finger(); | |
2965 if (task_finger != NULL && task_finger < _min_finger) | |
2966 _min_finger = task_finger; | |
2967 } | |
2968 | |
2969 _should_gray_objects = false; | |
2970 | |
2971 // This fixes a very subtle and fustrating bug. It might be the case | |
2972 // that, during en evacuation pause, heap regions that contain | |
2973 // objects that are gray (by being in regions contained in the | |
2974 // region stack) are included in the collection set. Since such gray | |
2975 // objects will be moved, and because it's not easy to redirect | |
2976 // region stack entries to point to a new location (because objects | |
2977 // in one region might be scattered to multiple regions after they | |
2978 // are copied), one option is to ensure that all marked objects | |
2979 // copied during a pause are pushed on the stack. Notice, however, | |
2980 // that this problem can only happen when the region stack is not | |
2981 // empty during an evacuation pause. So, we make the fix a bit less | |
2982 // conservative and ensure that regions are pushed on the stack, | |
2983 // irrespective whether all collection set regions are below the | |
2984 // finger, if the region stack is not empty. This is expected to be | |
2985 // a rare case, so I don't think it's necessary to be smarted about it. | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
2986 if (!region_stack_empty() || has_aborted_regions()) |
342 | 2987 _should_gray_objects = true; |
2988 } | |
2989 | |
2990 void ConcurrentMark::registerCSetRegion(HeapRegion* hr) { | |
2991 if (!concurrent_marking_in_progress()) | |
2992 return; | |
2993 | |
2994 HeapWord* region_end = hr->end(); | |
2995 if (region_end > _min_finger) | |
2996 _should_gray_objects = true; | |
2997 } | |
2998 | |
2999 // abandon current marking iteration due to a Full GC | |
3000 void ConcurrentMark::abort() { | |
3001 // Clear all marks to force marking thread to do nothing | |
3002 _nextMarkBitMap->clearAll(); | |
3003 // Empty mark stack | |
3004 clear_marking_state(); | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3005 for (int i = 0; i < (int)_max_task_num; ++i) { |
342 | 3006 _tasks[i]->clear_region_fields(); |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3007 } |
342 | 3008 _has_aborted = true; |
3009 | |
3010 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); | |
3011 satb_mq_set.abandon_partial_marking(); | |
1317
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
3012 // This can be called either during or outside marking, we'll read |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
3013 // the expected_active value from the SATB queue set. |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
3014 satb_mq_set.set_active_all_threads( |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
3015 false, /* new active value */ |
d4197f8d516a
6935821: G1: threads created during marking do not active their SATB queues
tonyp
parents:
1314
diff
changeset
|
3016 satb_mq_set.is_active() /* expected_active */); |
342 | 3017 } |
3018 | |
3019 static void print_ms_time_info(const char* prefix, const char* name, | |
3020 NumberSeq& ns) { | |
3021 gclog_or_tty->print_cr("%s%5d %12s: total time = %8.2f s (avg = %8.2f ms).", | |
3022 prefix, ns.num(), name, ns.sum()/1000.0, ns.avg()); | |
3023 if (ns.num() > 0) { | |
3024 gclog_or_tty->print_cr("%s [std. dev = %8.2f ms, max = %8.2f ms]", | |
3025 prefix, ns.sd(), ns.maximum()); | |
3026 } | |
3027 } | |
3028 | |
3029 void ConcurrentMark::print_summary_info() { | |
3030 gclog_or_tty->print_cr(" Concurrent marking:"); | |
3031 print_ms_time_info(" ", "init marks", _init_times); | |
3032 print_ms_time_info(" ", "remarks", _remark_times); | |
3033 { | |
3034 print_ms_time_info(" ", "final marks", _remark_mark_times); | |
3035 print_ms_time_info(" ", "weak refs", _remark_weak_ref_times); | |
3036 | |
3037 } | |
3038 print_ms_time_info(" ", "cleanups", _cleanup_times); | |
3039 gclog_or_tty->print_cr(" Final counting total time = %8.2f s (avg = %8.2f ms).", | |
3040 _total_counting_time, | |
3041 (_cleanup_times.num() > 0 ? _total_counting_time * 1000.0 / | |
3042 (double)_cleanup_times.num() | |
3043 : 0.0)); | |
3044 if (G1ScrubRemSets) { | |
3045 gclog_or_tty->print_cr(" RS scrub total time = %8.2f s (avg = %8.2f ms).", | |
3046 _total_rs_scrub_time, | |
3047 (_cleanup_times.num() > 0 ? _total_rs_scrub_time * 1000.0 / | |
3048 (double)_cleanup_times.num() | |
3049 : 0.0)); | |
3050 } | |
3051 gclog_or_tty->print_cr(" Total stop_world time = %8.2f s.", | |
3052 (_init_times.sum() + _remark_times.sum() + | |
3053 _cleanup_times.sum())/1000.0); | |
3054 gclog_or_tty->print_cr(" Total concurrent time = %8.2f s " | |
3055 "(%8.2f s marking, %8.2f s counting).", | |
3056 cmThread()->vtime_accum(), | |
3057 cmThread()->vtime_mark_accum(), | |
3058 cmThread()->vtime_count_accum()); | |
3059 } | |
3060 | |
1019 | 3061 void ConcurrentMark::print_worker_threads_on(outputStream* st) const { |
3062 _parallel_workers->print_worker_threads_on(st); | |
3063 } | |
3064 | |
342 | 3065 // Closures |
3066 // XXX: there seems to be a lot of code duplication here; | |
3067 // should refactor and consolidate the shared code. | |
3068 | |
3069 // This closure is used to mark refs into the CMS generation in | |
3070 // the CMS bit map. Called at the first checkpoint. | |
3071 | |
3072 // We take a break if someone is trying to stop the world. | |
3073 bool ConcurrentMark::do_yield_check(int worker_i) { | |
3074 if (should_yield()) { | |
3075 if (worker_i == 0) | |
3076 _g1h->g1_policy()->record_concurrent_pause(); | |
3077 cmThread()->yield(); | |
3078 if (worker_i == 0) | |
3079 _g1h->g1_policy()->record_concurrent_pause_end(); | |
3080 return true; | |
3081 } else { | |
3082 return false; | |
3083 } | |
3084 } | |
3085 | |
3086 bool ConcurrentMark::should_yield() { | |
3087 return cmThread()->should_yield(); | |
3088 } | |
3089 | |
3090 bool ConcurrentMark::containing_card_is_marked(void* p) { | |
3091 size_t offset = pointer_delta(p, _g1h->reserved_region().start(), 1); | |
3092 return _card_bm.at(offset >> CardTableModRefBS::card_shift); | |
3093 } | |
3094 | |
3095 bool ConcurrentMark::containing_cards_are_marked(void* start, | |
3096 void* last) { | |
3097 return | |
3098 containing_card_is_marked(start) && | |
3099 containing_card_is_marked(last); | |
3100 } | |
3101 | |
3102 #ifndef PRODUCT | |
3103 // for debugging purposes | |
3104 void ConcurrentMark::print_finger() { | |
3105 gclog_or_tty->print_cr("heap ["PTR_FORMAT", "PTR_FORMAT"), global finger = "PTR_FORMAT, | |
3106 _heap_start, _heap_end, _finger); | |
3107 for (int i = 0; i < (int) _max_task_num; ++i) { | |
3108 gclog_or_tty->print(" %d: "PTR_FORMAT, i, _tasks[i]->finger()); | |
3109 } | |
3110 gclog_or_tty->print_cr(""); | |
3111 } | |
3112 #endif | |
3113 | |
3114 // Closure for iteration over bitmaps | |
3115 class CMBitMapClosure : public BitMapClosure { | |
3116 private: | |
3117 // the bitmap that is being iterated over | |
3118 CMBitMap* _nextMarkBitMap; | |
3119 ConcurrentMark* _cm; | |
3120 CMTask* _task; | |
3121 // true if we're scanning a heap region claimed by the task (so that | |
3122 // we move the finger along), false if we're not, i.e. currently when | |
3123 // scanning a heap region popped from the region stack (so that we | |
3124 // do not move the task finger along; it'd be a mistake if we did so). | |
3125 bool _scanning_heap_region; | |
3126 | |
3127 public: | |
3128 CMBitMapClosure(CMTask *task, | |
3129 ConcurrentMark* cm, | |
3130 CMBitMap* nextMarkBitMap) | |
3131 : _task(task), _cm(cm), _nextMarkBitMap(nextMarkBitMap) { } | |
3132 | |
3133 void set_scanning_heap_region(bool scanning_heap_region) { | |
3134 _scanning_heap_region = scanning_heap_region; | |
3135 } | |
3136 | |
3137 bool do_bit(size_t offset) { | |
3138 HeapWord* addr = _nextMarkBitMap->offsetToHeapWord(offset); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3139 assert(_nextMarkBitMap->isMarked(addr), "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3140 assert( addr < _cm->finger(), "invariant"); |
342 | 3141 |
3142 if (_scanning_heap_region) { | |
3143 statsOnly( _task->increase_objs_found_on_bitmap() ); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3144 assert(addr >= _task->finger(), "invariant"); |
342 | 3145 // We move that task's local finger along. |
3146 _task->move_finger_to(addr); | |
3147 } else { | |
3148 // We move the task's region finger along. | |
3149 _task->move_region_finger_to(addr); | |
3150 } | |
3151 | |
3152 _task->scan_object(oop(addr)); | |
3153 // we only partially drain the local queue and global stack | |
3154 _task->drain_local_queue(true); | |
3155 _task->drain_global_stack(true); | |
3156 | |
3157 // if the has_aborted flag has been raised, we need to bail out of | |
3158 // the iteration | |
3159 return !_task->has_aborted(); | |
3160 } | |
3161 }; | |
3162 | |
3163 // Closure for iterating over objects, currently only used for | |
3164 // processing SATB buffers. | |
3165 class CMObjectClosure : public ObjectClosure { | |
3166 private: | |
3167 CMTask* _task; | |
3168 | |
3169 public: | |
3170 void do_object(oop obj) { | |
3171 _task->deal_with_reference(obj); | |
3172 } | |
3173 | |
3174 CMObjectClosure(CMTask* task) : _task(task) { } | |
3175 }; | |
3176 | |
3177 // Closure for iterating over object fields | |
3178 class CMOopClosure : public OopClosure { | |
3179 private: | |
3180 G1CollectedHeap* _g1h; | |
3181 ConcurrentMark* _cm; | |
3182 CMTask* _task; | |
3183 | |
3184 public: | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3185 virtual void do_oop(narrowOop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3186 virtual void do_oop( oop* p) { do_oop_work(p); } |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3187 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3188 template <class T> void do_oop_work(T* p) { |
2152 | 3189 assert( _g1h->is_in_g1_reserved((HeapWord*) p), "invariant"); |
2361 | 3190 assert(!_g1h->is_on_master_free_list( |
2152 | 3191 _g1h->heap_region_containing((HeapWord*) p)), "invariant"); |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3192 |
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3193 oop obj = oopDesc::load_decode_heap_oop(p); |
342 | 3194 if (_cm->verbose_high()) |
3195 gclog_or_tty->print_cr("[%d] we're looking at location " | |
3196 "*"PTR_FORMAT" = "PTR_FORMAT, | |
3197 _task->task_id(), p, (void*) obj); | |
3198 _task->deal_with_reference(obj); | |
3199 } | |
3200 | |
3201 CMOopClosure(G1CollectedHeap* g1h, | |
3202 ConcurrentMark* cm, | |
3203 CMTask* task) | |
1974
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
3204 : _g1h(g1h), _cm(cm), _task(task) |
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
3205 { |
2436 | 3206 assert(_ref_processor == NULL, "should be initialized to NULL"); |
3207 | |
3208 if (G1UseConcMarkReferenceProcessing) { | |
3209 _ref_processor = g1h->ref_processor(); | |
3210 assert(_ref_processor != NULL, "should not be NULL"); | |
3211 } | |
1974
fd1d227ef1b9
6983204: G1: Nightly test nsk/regression/b4958615 failing with +ExplicitGCInvokesConcurrent
johnc
parents:
1972
diff
changeset
|
3212 } |
342 | 3213 }; |
3214 | |
3215 void CMTask::setup_for_region(HeapRegion* hr) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3216 // Separated the asserts so that we know which one fires. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3217 assert(hr != NULL, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3218 "claim_region() should have filtered out continues humongous regions"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3219 assert(!hr->continuesHumongous(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3220 "claim_region() should have filtered out continues humongous regions"); |
342 | 3221 |
3222 if (_cm->verbose_low()) | |
3223 gclog_or_tty->print_cr("[%d] setting up for region "PTR_FORMAT, | |
3224 _task_id, hr); | |
3225 | |
3226 _curr_region = hr; | |
3227 _finger = hr->bottom(); | |
3228 update_region_limit(); | |
3229 } | |
3230 | |
3231 void CMTask::update_region_limit() { | |
3232 HeapRegion* hr = _curr_region; | |
3233 HeapWord* bottom = hr->bottom(); | |
3234 HeapWord* limit = hr->next_top_at_mark_start(); | |
3235 | |
3236 if (limit == bottom) { | |
3237 if (_cm->verbose_low()) | |
3238 gclog_or_tty->print_cr("[%d] found an empty region " | |
3239 "["PTR_FORMAT", "PTR_FORMAT")", | |
3240 _task_id, bottom, limit); | |
3241 // The region was collected underneath our feet. | |
3242 // We set the finger to bottom to ensure that the bitmap | |
3243 // iteration that will follow this will not do anything. | |
3244 // (this is not a condition that holds when we set the region up, | |
3245 // as the region is not supposed to be empty in the first place) | |
3246 _finger = bottom; | |
3247 } else if (limit >= _region_limit) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3248 assert(limit >= _finger, "peace of mind"); |
342 | 3249 } else { |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3250 assert(limit < _region_limit, "only way to get here"); |
342 | 3251 // This can happen under some pretty unusual circumstances. An |
3252 // evacuation pause empties the region underneath our feet (NTAMS | |
3253 // at bottom). We then do some allocation in the region (NTAMS | |
3254 // stays at bottom), followed by the region being used as a GC | |
3255 // alloc region (NTAMS will move to top() and the objects | |
3256 // originally below it will be grayed). All objects now marked in | |
3257 // the region are explicitly grayed, if below the global finger, | |
3258 // and we do not need in fact to scan anything else. So, we simply | |
3259 // set _finger to be limit to ensure that the bitmap iteration | |
3260 // doesn't do anything. | |
3261 _finger = limit; | |
3262 } | |
3263 | |
3264 _region_limit = limit; | |
3265 } | |
3266 | |
3267 void CMTask::giveup_current_region() { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3268 assert(_curr_region != NULL, "invariant"); |
342 | 3269 if (_cm->verbose_low()) |
3270 gclog_or_tty->print_cr("[%d] giving up region "PTR_FORMAT, | |
3271 _task_id, _curr_region); | |
3272 clear_region_fields(); | |
3273 } | |
3274 | |
3275 void CMTask::clear_region_fields() { | |
3276 // Values for these three fields that indicate that we're not | |
3277 // holding on to a region. | |
3278 _curr_region = NULL; | |
3279 _finger = NULL; | |
3280 _region_limit = NULL; | |
3281 | |
3282 _region_finger = NULL; | |
3283 } | |
3284 | |
3285 void CMTask::reset(CMBitMap* nextMarkBitMap) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3286 guarantee(nextMarkBitMap != NULL, "invariant"); |
342 | 3287 |
3288 if (_cm->verbose_low()) | |
3289 gclog_or_tty->print_cr("[%d] resetting", _task_id); | |
3290 | |
3291 _nextMarkBitMap = nextMarkBitMap; | |
3292 clear_region_fields(); | |
1885
a5c514e74487
6988458: G1: assert(mr.end() <= _cm->finger()) failed: otherwise the region shouldn't be on the stack
johnc
parents:
1835
diff
changeset
|
3293 assert(_aborted_region.is_empty(), "should have been cleared"); |
342 | 3294 |
3295 _calls = 0; | |
3296 _elapsed_time_ms = 0.0; | |
3297 _termination_time_ms = 0.0; | |
3298 _termination_start_time_ms = 0.0; | |
3299 | |
3300 #if _MARKING_STATS_ | |
3301 _local_pushes = 0; | |
3302 _local_pops = 0; | |
3303 _local_max_size = 0; | |
3304 _objs_scanned = 0; | |
3305 _global_pushes = 0; | |
3306 _global_pops = 0; | |
3307 _global_max_size = 0; | |
3308 _global_transfers_to = 0; | |
3309 _global_transfers_from = 0; | |
3310 _region_stack_pops = 0; | |
3311 _regions_claimed = 0; | |
3312 _objs_found_on_bitmap = 0; | |
3313 _satb_buffers_processed = 0; | |
3314 _steal_attempts = 0; | |
3315 _steals = 0; | |
3316 _aborted = 0; | |
3317 _aborted_overflow = 0; | |
3318 _aborted_cm_aborted = 0; | |
3319 _aborted_yield = 0; | |
3320 _aborted_timed_out = 0; | |
3321 _aborted_satb = 0; | |
3322 _aborted_termination = 0; | |
3323 #endif // _MARKING_STATS_ | |
3324 } | |
3325 | |
3326 bool CMTask::should_exit_termination() { | |
3327 regular_clock_call(); | |
3328 // This is called when we are in the termination protocol. We should | |
3329 // quit if, for some reason, this task wants to abort or the global | |
3330 // stack is not empty (this means that we can get work from it). | |
3331 return !_cm->mark_stack_empty() || has_aborted(); | |
3332 } | |
3333 | |
3334 // This determines whether the method below will check both the local | |
3335 // and global fingers when determining whether to push on the stack a | |
3336 // gray object (value 1) or whether it will only check the global one | |
3337 // (value 0). The tradeoffs are that the former will be a bit more | |
3338 // accurate and possibly push less on the stack, but it might also be | |
3339 // a little bit slower. | |
3340 | |
3341 #define _CHECK_BOTH_FINGERS_ 1 | |
3342 | |
3343 void CMTask::deal_with_reference(oop obj) { | |
3344 if (_cm->verbose_high()) | |
3345 gclog_or_tty->print_cr("[%d] we're dealing with reference = "PTR_FORMAT, | |
3346 _task_id, (void*) obj); | |
3347 | |
3348 ++_refs_reached; | |
3349 | |
3350 HeapWord* objAddr = (HeapWord*) obj; | |
845
df6caf649ff7
6700789: G1: Enable use of compressed oops with G1 heaps
ysr
parents:
831
diff
changeset
|
3351 assert(obj->is_oop_or_null(true /* ignore mark word */), "Error"); |
342 | 3352 if (_g1h->is_in_g1_reserved(objAddr)) { |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3353 assert(obj != NULL, "is_in_g1_reserved should ensure this"); |
342 | 3354 HeapRegion* hr = _g1h->heap_region_containing(obj); |
3355 if (_g1h->is_obj_ill(obj, hr)) { | |
3356 if (_cm->verbose_high()) | |
3357 gclog_or_tty->print_cr("[%d] "PTR_FORMAT" is not considered marked", | |
3358 _task_id, (void*) obj); | |
3359 | |
3360 // we need to mark it first | |
3361 if (_nextMarkBitMap->parMark(objAddr)) { | |
3362 // No OrderAccess:store_load() is needed. It is implicit in the | |
3363 // CAS done in parMark(objAddr) above | |
3364 HeapWord* global_finger = _cm->finger(); | |
3365 | |
3366 #if _CHECK_BOTH_FINGERS_ | |
3367 // we will check both the local and global fingers | |
3368 | |
3369 if (_finger != NULL && objAddr < _finger) { | |
3370 if (_cm->verbose_high()) | |
3371 gclog_or_tty->print_cr("[%d] below the local finger ("PTR_FORMAT"), " | |
3372 "pushing it", _task_id, _finger); | |
3373 push(obj); | |
3374 } else if (_curr_region != NULL && objAddr < _region_limit) { | |
3375 // do nothing | |
3376 } else if (objAddr < global_finger) { | |
3377 // Notice that the global finger might be moving forward | |
3378 // concurrently. This is not a problem. In the worst case, we | |
3379 // mark the object while it is above the global finger and, by | |
3380 // the time we read the global finger, it has moved forward | |
3381 // passed this object. In this case, the object will probably | |
3382 // be visited when a task is scanning the region and will also | |
3383 // be pushed on the stack. So, some duplicate work, but no | |
3384 // correctness problems. | |
3385 | |
3386 if (_cm->verbose_high()) | |
3387 gclog_or_tty->print_cr("[%d] below the global finger " | |
3388 "("PTR_FORMAT"), pushing it", | |
3389 _task_id, global_finger); | |
3390 push(obj); | |
3391 } else { | |
3392 // do nothing | |
3393 } | |
3394 #else // _CHECK_BOTH_FINGERS_ | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
3395 // we will only check the global finger |
342 | 3396 |
3397 if (objAddr < global_finger) { | |
3398 // see long comment above | |
3399 | |
3400 if (_cm->verbose_high()) | |
3401 gclog_or_tty->print_cr("[%d] below the global finger " | |
3402 "("PTR_FORMAT"), pushing it", | |
3403 _task_id, global_finger); | |
3404 push(obj); | |
3405 } | |
3406 #endif // _CHECK_BOTH_FINGERS_ | |
3407 } | |
3408 } | |
3409 } | |
3410 } | |
3411 | |
3412 void CMTask::push(oop obj) { | |
3413 HeapWord* objAddr = (HeapWord*) obj; | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3414 assert(_g1h->is_in_g1_reserved(objAddr), "invariant"); |
2361 | 3415 assert(!_g1h->is_on_master_free_list( |
2152 | 3416 _g1h->heap_region_containing((HeapWord*) objAddr)), "invariant"); |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3417 assert(!_g1h->is_obj_ill(obj), "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3418 assert(_nextMarkBitMap->isMarked(objAddr), "invariant"); |
342 | 3419 |
3420 if (_cm->verbose_high()) | |
3421 gclog_or_tty->print_cr("[%d] pushing "PTR_FORMAT, _task_id, (void*) obj); | |
3422 | |
3423 if (!_task_queue->push(obj)) { | |
3424 // The local task queue looks full. We need to push some entries | |
3425 // to the global stack. | |
3426 | |
3427 if (_cm->verbose_medium()) | |
3428 gclog_or_tty->print_cr("[%d] task queue overflow, " | |
3429 "moving entries to the global stack", | |
3430 _task_id); | |
3431 move_entries_to_global_stack(); | |
3432 | |
3433 // this should succeed since, even if we overflow the global | |
3434 // stack, we should have definitely removed some entries from the | |
3435 // local queue. So, there must be space on it. | |
3436 bool success = _task_queue->push(obj); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3437 assert(success, "invariant"); |
342 | 3438 } |
3439 | |
3440 statsOnly( int tmp_size = _task_queue->size(); | |
3441 if (tmp_size > _local_max_size) | |
3442 _local_max_size = tmp_size; | |
3443 ++_local_pushes ); | |
3444 } | |
3445 | |
3446 void CMTask::reached_limit() { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3447 assert(_words_scanned >= _words_scanned_limit || |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3448 _refs_reached >= _refs_reached_limit , |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3449 "shouldn't have been called otherwise"); |
342 | 3450 regular_clock_call(); |
3451 } | |
3452 | |
3453 void CMTask::regular_clock_call() { | |
3454 if (has_aborted()) | |
3455 return; | |
3456 | |
3457 // First, we need to recalculate the words scanned and refs reached | |
3458 // limits for the next clock call. | |
3459 recalculate_limits(); | |
3460 | |
3461 // During the regular clock call we do the following | |
3462 | |
3463 // (1) If an overflow has been flagged, then we abort. | |
3464 if (_cm->has_overflown()) { | |
3465 set_has_aborted(); | |
3466 return; | |
3467 } | |
3468 | |
3469 // If we are not concurrent (i.e. we're doing remark) we don't need | |
3470 // to check anything else. The other steps are only needed during | |
3471 // the concurrent marking phase. | |
3472 if (!concurrent()) | |
3473 return; | |
3474 | |
3475 // (2) If marking has been aborted for Full GC, then we also abort. | |
3476 if (_cm->has_aborted()) { | |
3477 set_has_aborted(); | |
3478 statsOnly( ++_aborted_cm_aborted ); | |
3479 return; | |
3480 } | |
3481 | |
3482 double curr_time_ms = os::elapsedVTime() * 1000.0; | |
3483 | |
3484 // (3) If marking stats are enabled, then we update the step history. | |
3485 #if _MARKING_STATS_ | |
3486 if (_words_scanned >= _words_scanned_limit) | |
3487 ++_clock_due_to_scanning; | |
3488 if (_refs_reached >= _refs_reached_limit) | |
3489 ++_clock_due_to_marking; | |
3490 | |
3491 double last_interval_ms = curr_time_ms - _interval_start_time_ms; | |
3492 _interval_start_time_ms = curr_time_ms; | |
3493 _all_clock_intervals_ms.add(last_interval_ms); | |
3494 | |
3495 if (_cm->verbose_medium()) { | |
3496 gclog_or_tty->print_cr("[%d] regular clock, interval = %1.2lfms, " | |
3497 "scanned = %d%s, refs reached = %d%s", | |
3498 _task_id, last_interval_ms, | |
3499 _words_scanned, | |
3500 (_words_scanned >= _words_scanned_limit) ? " (*)" : "", | |
3501 _refs_reached, | |
3502 (_refs_reached >= _refs_reached_limit) ? " (*)" : ""); | |
3503 } | |
3504 #endif // _MARKING_STATS_ | |
3505 | |
3506 // (4) We check whether we should yield. If we have to, then we abort. | |
3507 if (_cm->should_yield()) { | |
3508 // We should yield. To do this we abort the task. The caller is | |
3509 // responsible for yielding. | |
3510 set_has_aborted(); | |
3511 statsOnly( ++_aborted_yield ); | |
3512 return; | |
3513 } | |
3514 | |
3515 // (5) We check whether we've reached our time quota. If we have, | |
3516 // then we abort. | |
3517 double elapsed_time_ms = curr_time_ms - _start_time_ms; | |
3518 if (elapsed_time_ms > _time_target_ms) { | |
3519 set_has_aborted(); | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
3520 _has_timed_out = true; |
342 | 3521 statsOnly( ++_aborted_timed_out ); |
3522 return; | |
3523 } | |
3524 | |
3525 // (6) Finally, we check whether there are enough completed STAB | |
3526 // buffers available for processing. If there are, we abort. | |
3527 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); | |
3528 if (!_draining_satb_buffers && satb_mq_set.process_completed_buffers()) { | |
3529 if (_cm->verbose_low()) | |
3530 gclog_or_tty->print_cr("[%d] aborting to deal with pending SATB buffers", | |
3531 _task_id); | |
3532 // we do need to process SATB buffers, we'll abort and restart | |
3533 // the marking task to do so | |
3534 set_has_aborted(); | |
3535 statsOnly( ++_aborted_satb ); | |
3536 return; | |
3537 } | |
3538 } | |
3539 | |
3540 void CMTask::recalculate_limits() { | |
3541 _real_words_scanned_limit = _words_scanned + words_scanned_period; | |
3542 _words_scanned_limit = _real_words_scanned_limit; | |
3543 | |
3544 _real_refs_reached_limit = _refs_reached + refs_reached_period; | |
3545 _refs_reached_limit = _real_refs_reached_limit; | |
3546 } | |
3547 | |
3548 void CMTask::decrease_limits() { | |
3549 // This is called when we believe that we're going to do an infrequent | |
3550 // operation which will increase the per byte scanned cost (i.e. move | |
3551 // entries to/from the global stack). It basically tries to decrease the | |
3552 // scanning limit so that the clock is called earlier. | |
3553 | |
3554 if (_cm->verbose_medium()) | |
3555 gclog_or_tty->print_cr("[%d] decreasing limits", _task_id); | |
3556 | |
3557 _words_scanned_limit = _real_words_scanned_limit - | |
3558 3 * words_scanned_period / 4; | |
3559 _refs_reached_limit = _real_refs_reached_limit - | |
3560 3 * refs_reached_period / 4; | |
3561 } | |
3562 | |
3563 void CMTask::move_entries_to_global_stack() { | |
3564 // local array where we'll store the entries that will be popped | |
3565 // from the local queue | |
3566 oop buffer[global_stack_transfer_size]; | |
3567 | |
3568 int n = 0; | |
3569 oop obj; | |
3570 while (n < global_stack_transfer_size && _task_queue->pop_local(obj)) { | |
3571 buffer[n] = obj; | |
3572 ++n; | |
3573 } | |
3574 | |
3575 if (n > 0) { | |
3576 // we popped at least one entry from the local queue | |
3577 | |
3578 statsOnly( ++_global_transfers_to; _local_pops += n ); | |
3579 | |
3580 if (!_cm->mark_stack_push(buffer, n)) { | |
3581 if (_cm->verbose_low()) | |
3582 gclog_or_tty->print_cr("[%d] aborting due to global stack overflow", _task_id); | |
3583 set_has_aborted(); | |
3584 } else { | |
3585 // the transfer was successful | |
3586 | |
3587 if (_cm->verbose_medium()) | |
3588 gclog_or_tty->print_cr("[%d] pushed %d entries to the global stack", | |
3589 _task_id, n); | |
3590 statsOnly( int tmp_size = _cm->mark_stack_size(); | |
3591 if (tmp_size > _global_max_size) | |
3592 _global_max_size = tmp_size; | |
3593 _global_pushes += n ); | |
3594 } | |
3595 } | |
3596 | |
3597 // this operation was quite expensive, so decrease the limits | |
3598 decrease_limits(); | |
3599 } | |
3600 | |
3601 void CMTask::get_entries_from_global_stack() { | |
3602 // local array where we'll store the entries that will be popped | |
3603 // from the global stack. | |
3604 oop buffer[global_stack_transfer_size]; | |
3605 int n; | |
3606 _cm->mark_stack_pop(buffer, global_stack_transfer_size, &n); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3607 assert(n <= global_stack_transfer_size, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3608 "we should not pop more than the given limit"); |
342 | 3609 if (n > 0) { |
3610 // yes, we did actually pop at least one entry | |
3611 | |
3612 statsOnly( ++_global_transfers_from; _global_pops += n ); | |
3613 if (_cm->verbose_medium()) | |
3614 gclog_or_tty->print_cr("[%d] popped %d entries from the global stack", | |
3615 _task_id, n); | |
3616 for (int i = 0; i < n; ++i) { | |
3617 bool success = _task_queue->push(buffer[i]); | |
3618 // We only call this when the local queue is empty or under a | |
3619 // given target limit. So, we do not expect this push to fail. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3620 assert(success, "invariant"); |
342 | 3621 } |
3622 | |
3623 statsOnly( int tmp_size = _task_queue->size(); | |
3624 if (tmp_size > _local_max_size) | |
3625 _local_max_size = tmp_size; | |
3626 _local_pushes += n ); | |
3627 } | |
3628 | |
3629 // this operation was quite expensive, so decrease the limits | |
3630 decrease_limits(); | |
3631 } | |
3632 | |
3633 void CMTask::drain_local_queue(bool partially) { | |
3634 if (has_aborted()) | |
3635 return; | |
3636 | |
3637 // Decide what the target size is, depending whether we're going to | |
3638 // drain it partially (so that other tasks can steal if they run out | |
3639 // of things to do) or totally (at the very end). | |
3640 size_t target_size; | |
3641 if (partially) | |
3642 target_size = MIN2((size_t)_task_queue->max_elems()/3, GCDrainStackTargetSize); | |
3643 else | |
3644 target_size = 0; | |
3645 | |
3646 if (_task_queue->size() > target_size) { | |
3647 if (_cm->verbose_high()) | |
3648 gclog_or_tty->print_cr("[%d] draining local queue, target size = %d", | |
3649 _task_id, target_size); | |
3650 | |
3651 oop obj; | |
3652 bool ret = _task_queue->pop_local(obj); | |
3653 while (ret) { | |
3654 statsOnly( ++_local_pops ); | |
3655 | |
3656 if (_cm->verbose_high()) | |
3657 gclog_or_tty->print_cr("[%d] popped "PTR_FORMAT, _task_id, | |
3658 (void*) obj); | |
3659 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3660 assert(_g1h->is_in_g1_reserved((HeapWord*) obj), "invariant" ); |
2361 | 3661 assert(!_g1h->is_on_master_free_list( |
2152 | 3662 _g1h->heap_region_containing((HeapWord*) obj)), "invariant"); |
342 | 3663 |
3664 scan_object(obj); | |
3665 | |
3666 if (_task_queue->size() <= target_size || has_aborted()) | |
3667 ret = false; | |
3668 else | |
3669 ret = _task_queue->pop_local(obj); | |
3670 } | |
3671 | |
3672 if (_cm->verbose_high()) | |
3673 gclog_or_tty->print_cr("[%d] drained local queue, size = %d", | |
3674 _task_id, _task_queue->size()); | |
3675 } | |
3676 } | |
3677 | |
3678 void CMTask::drain_global_stack(bool partially) { | |
3679 if (has_aborted()) | |
3680 return; | |
3681 | |
3682 // We have a policy to drain the local queue before we attempt to | |
3683 // drain the global stack. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3684 assert(partially || _task_queue->size() == 0, "invariant"); |
342 | 3685 |
3686 // Decide what the target size is, depending whether we're going to | |
3687 // drain it partially (so that other tasks can steal if they run out | |
3688 // of things to do) or totally (at the very end). Notice that, | |
3689 // because we move entries from the global stack in chunks or | |
3690 // because another task might be doing the same, we might in fact | |
3691 // drop below the target. But, this is not a problem. | |
3692 size_t target_size; | |
3693 if (partially) | |
3694 target_size = _cm->partial_mark_stack_size_target(); | |
3695 else | |
3696 target_size = 0; | |
3697 | |
3698 if (_cm->mark_stack_size() > target_size) { | |
3699 if (_cm->verbose_low()) | |
3700 gclog_or_tty->print_cr("[%d] draining global_stack, target size %d", | |
3701 _task_id, target_size); | |
3702 | |
3703 while (!has_aborted() && _cm->mark_stack_size() > target_size) { | |
3704 get_entries_from_global_stack(); | |
3705 drain_local_queue(partially); | |
3706 } | |
3707 | |
3708 if (_cm->verbose_low()) | |
3709 gclog_or_tty->print_cr("[%d] drained global stack, size = %d", | |
3710 _task_id, _cm->mark_stack_size()); | |
3711 } | |
3712 } | |
3713 | |
3714 // SATB Queue has several assumptions on whether to call the par or | |
3715 // non-par versions of the methods. this is why some of the code is | |
3716 // replicated. We should really get rid of the single-threaded version | |
3717 // of the code to simplify things. | |
3718 void CMTask::drain_satb_buffers() { | |
3719 if (has_aborted()) | |
3720 return; | |
3721 | |
3722 // We set this so that the regular clock knows that we're in the | |
3723 // middle of draining buffers and doesn't set the abort flag when it | |
3724 // notices that SATB buffers are available for draining. It'd be | |
3725 // very counter productive if it did that. :-) | |
3726 _draining_satb_buffers = true; | |
3727 | |
3728 CMObjectClosure oc(this); | |
3729 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
3730 if (G1CollectedHeap::use_parallel_gc_threads()) |
342 | 3731 satb_mq_set.set_par_closure(_task_id, &oc); |
3732 else | |
3733 satb_mq_set.set_closure(&oc); | |
3734 | |
3735 // This keeps claiming and applying the closure to completed buffers | |
3736 // until we run out of buffers or we need to abort. | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
3737 if (G1CollectedHeap::use_parallel_gc_threads()) { |
342 | 3738 while (!has_aborted() && |
3739 satb_mq_set.par_apply_closure_to_completed_buffer(_task_id)) { | |
3740 if (_cm->verbose_medium()) | |
3741 gclog_or_tty->print_cr("[%d] processed an SATB buffer", _task_id); | |
3742 statsOnly( ++_satb_buffers_processed ); | |
3743 regular_clock_call(); | |
3744 } | |
3745 } else { | |
3746 while (!has_aborted() && | |
3747 satb_mq_set.apply_closure_to_completed_buffer()) { | |
3748 if (_cm->verbose_medium()) | |
3749 gclog_or_tty->print_cr("[%d] processed an SATB buffer", _task_id); | |
3750 statsOnly( ++_satb_buffers_processed ); | |
3751 regular_clock_call(); | |
3752 } | |
3753 } | |
3754 | |
3755 if (!concurrent() && !has_aborted()) { | |
3756 // We should only do this during remark. | |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
3757 if (G1CollectedHeap::use_parallel_gc_threads()) |
342 | 3758 satb_mq_set.par_iterate_closure_all_threads(_task_id); |
3759 else | |
3760 satb_mq_set.iterate_closure_all_threads(); | |
3761 } | |
3762 | |
3763 _draining_satb_buffers = false; | |
3764 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3765 assert(has_aborted() || |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3766 concurrent() || |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3767 satb_mq_set.completed_buffers_num() == 0, "invariant"); |
342 | 3768 |
1833
8b10f48633dc
6984287: Regularize how GC parallel workers are specified.
jmasa
parents:
1719
diff
changeset
|
3769 if (G1CollectedHeap::use_parallel_gc_threads()) |
342 | 3770 satb_mq_set.set_par_closure(_task_id, NULL); |
3771 else | |
3772 satb_mq_set.set_closure(NULL); | |
3773 | |
3774 // again, this was a potentially expensive operation, decrease the | |
3775 // limits to get the regular clock call early | |
3776 decrease_limits(); | |
3777 } | |
3778 | |
3779 void CMTask::drain_region_stack(BitMapClosure* bc) { | |
3780 if (has_aborted()) | |
3781 return; | |
3782 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3783 assert(_region_finger == NULL, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3784 "it should be NULL when we're not scanning a region"); |
342 | 3785 |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3786 if (!_cm->region_stack_empty() || !_aborted_region.is_empty()) { |
342 | 3787 if (_cm->verbose_low()) |
3788 gclog_or_tty->print_cr("[%d] draining region stack, size = %d", | |
3789 _task_id, _cm->region_stack_size()); | |
3790 | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3791 MemRegion mr; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3792 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3793 if (!_aborted_region.is_empty()) { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3794 mr = _aborted_region; |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3795 _aborted_region = MemRegion(); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3796 |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3797 if (_cm->verbose_low()) |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3798 gclog_or_tty->print_cr("[%d] scanning aborted region [ " PTR_FORMAT ", " PTR_FORMAT " )", |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3799 _task_id, mr.start(), mr.end()); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3800 } else { |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3801 mr = _cm->region_stack_pop_lock_free(); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3802 // it returns MemRegion() if the pop fails |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3803 statsOnly(if (mr.start() != NULL) ++_region_stack_pops ); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3804 } |
342 | 3805 |
3806 while (mr.start() != NULL) { | |
3807 if (_cm->verbose_medium()) | |
3808 gclog_or_tty->print_cr("[%d] we are scanning region " | |
3809 "["PTR_FORMAT", "PTR_FORMAT")", | |
3810 _task_id, mr.start(), mr.end()); | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3811 |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3812 assert(mr.end() <= _cm->finger(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3813 "otherwise the region shouldn't be on the stack"); |
342 | 3814 assert(!mr.is_empty(), "Only non-empty regions live on the region stack"); |
3815 if (_nextMarkBitMap->iterate(bc, mr)) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3816 assert(!has_aborted(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3817 "cannot abort the task without aborting the bitmap iteration"); |
342 | 3818 |
3819 // We finished iterating over the region without aborting. | |
3820 regular_clock_call(); | |
3821 if (has_aborted()) | |
3822 mr = MemRegion(); | |
3823 else { | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3824 mr = _cm->region_stack_pop_lock_free(); |
342 | 3825 // it returns MemRegion() if the pop fails |
3826 statsOnly(if (mr.start() != NULL) ++_region_stack_pops ); | |
3827 } | |
3828 } else { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3829 assert(has_aborted(), "currently the only way to do so"); |
342 | 3830 |
3831 // The only way to abort the bitmap iteration is to return | |
3832 // false from the do_bit() method. However, inside the | |
3833 // do_bit() method we move the _region_finger to point to the | |
3834 // object currently being looked at. So, if we bail out, we | |
3835 // have definitely set _region_finger to something non-null. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
3836 assert(_region_finger != NULL, "invariant"); |
342 | 3837 |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3838 // Make sure that any previously aborted region has been |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3839 // cleared. |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3840 assert(_aborted_region.is_empty(), "aborted region not cleared"); |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3841 |
342 | 3842 // The iteration was actually aborted. So now _region_finger |
3843 // points to the address of the object we last scanned. If we | |
3844 // leave it there, when we restart this task, we will rescan | |
3845 // the object. It is easy to avoid this. We move the finger by | |
3846 // enough to point to the next possible object header (the | |
3847 // bitmap knows by how much we need to move it as it knows its | |
3848 // granularity). | |
3849 MemRegion newRegion = | |
3850 MemRegion(_nextMarkBitMap->nextWord(_region_finger), mr.end()); | |
3851 | |
3852 if (!newRegion.is_empty()) { | |
3853 if (_cm->verbose_low()) { | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3854 gclog_or_tty->print_cr("[%d] recording unscanned region" |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3855 "[" PTR_FORMAT "," PTR_FORMAT ") in CMTask", |
342 | 3856 _task_id, |
3857 newRegion.start(), newRegion.end()); | |
3858 } | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3859 // Now record the part of the region we didn't scan to |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3860 // make sure this task scans it later. |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
3861 _aborted_region = newRegion; |
342 | 3862 } |
3863 // break from while | |
3864 mr = MemRegion(); | |
3865 } | |
3866 _region_finger = NULL; | |
3867 } | |
3868 | |
3869 if (_cm->verbose_low()) | |
3870 gclog_or_tty->print_cr("[%d] drained region stack, size = %d", | |
3871 _task_id, _cm->region_stack_size()); | |
3872 } | |
3873 } | |
3874 | |
3875 void CMTask::print_stats() { | |
3876 gclog_or_tty->print_cr("Marking Stats, task = %d, calls = %d", | |
3877 _task_id, _calls); | |
3878 gclog_or_tty->print_cr(" Elapsed time = %1.2lfms, Termination time = %1.2lfms", | |
3879 _elapsed_time_ms, _termination_time_ms); | |
3880 gclog_or_tty->print_cr(" Step Times (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms", | |
3881 _step_times_ms.num(), _step_times_ms.avg(), | |
3882 _step_times_ms.sd()); | |
3883 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms", | |
3884 _step_times_ms.maximum(), _step_times_ms.sum()); | |
3885 | |
3886 #if _MARKING_STATS_ | |
3887 gclog_or_tty->print_cr(" Clock Intervals (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms", | |
3888 _all_clock_intervals_ms.num(), _all_clock_intervals_ms.avg(), | |
3889 _all_clock_intervals_ms.sd()); | |
3890 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms", | |
3891 _all_clock_intervals_ms.maximum(), | |
3892 _all_clock_intervals_ms.sum()); | |
3893 gclog_or_tty->print_cr(" Clock Causes (cum): scanning = %d, marking = %d", | |
3894 _clock_due_to_scanning, _clock_due_to_marking); | |
3895 gclog_or_tty->print_cr(" Objects: scanned = %d, found on the bitmap = %d", | |
3896 _objs_scanned, _objs_found_on_bitmap); | |
3897 gclog_or_tty->print_cr(" Local Queue: pushes = %d, pops = %d, max size = %d", | |
3898 _local_pushes, _local_pops, _local_max_size); | |
3899 gclog_or_tty->print_cr(" Global Stack: pushes = %d, pops = %d, max size = %d", | |
3900 _global_pushes, _global_pops, _global_max_size); | |
3901 gclog_or_tty->print_cr(" transfers to = %d, transfers from = %d", | |
3902 _global_transfers_to,_global_transfers_from); | |
3903 gclog_or_tty->print_cr(" Regions: claimed = %d, Region Stack: pops = %d", | |
3904 _regions_claimed, _region_stack_pops); | |
3905 gclog_or_tty->print_cr(" SATB buffers: processed = %d", _satb_buffers_processed); | |
3906 gclog_or_tty->print_cr(" Steals: attempts = %d, successes = %d", | |
3907 _steal_attempts, _steals); | |
3908 gclog_or_tty->print_cr(" Aborted: %d, due to", _aborted); | |
3909 gclog_or_tty->print_cr(" overflow: %d, global abort: %d, yield: %d", | |
3910 _aborted_overflow, _aborted_cm_aborted, _aborted_yield); | |
3911 gclog_or_tty->print_cr(" time out: %d, SATB: %d, termination: %d", | |
3912 _aborted_timed_out, _aborted_satb, _aborted_termination); | |
3913 #endif // _MARKING_STATS_ | |
3914 } | |
3915 | |
3916 /***************************************************************************** | |
3917 | |
3918 The do_marking_step(time_target_ms) method is the building block | |
3919 of the parallel marking framework. It can be called in parallel | |
3920 with other invocations of do_marking_step() on different tasks | |
3921 (but only one per task, obviously) and concurrently with the | |
3922 mutator threads, or during remark, hence it eliminates the need | |
3923 for two versions of the code. When called during remark, it will | |
3924 pick up from where the task left off during the concurrent marking | |
3925 phase. Interestingly, tasks are also claimable during evacuation | |
3926 pauses too, since do_marking_step() ensures that it aborts before | |
3927 it needs to yield. | |
3928 | |
3929 The data structures that is uses to do marking work are the | |
3930 following: | |
3931 | |
3932 (1) Marking Bitmap. If there are gray objects that appear only | |
3933 on the bitmap (this happens either when dealing with an overflow | |
3934 or when the initial marking phase has simply marked the roots | |
3935 and didn't push them on the stack), then tasks claim heap | |
3936 regions whose bitmap they then scan to find gray objects. A | |
3937 global finger indicates where the end of the last claimed region | |
3938 is. A local finger indicates how far into the region a task has | |
3939 scanned. The two fingers are used to determine how to gray an | |
3940 object (i.e. whether simply marking it is OK, as it will be | |
3941 visited by a task in the future, or whether it needs to be also | |
3942 pushed on a stack). | |
3943 | |
3944 (2) Local Queue. The local queue of the task which is accessed | |
3945 reasonably efficiently by the task. Other tasks can steal from | |
3946 it when they run out of work. Throughout the marking phase, a | |
3947 task attempts to keep its local queue short but not totally | |
3948 empty, so that entries are available for stealing by other | |
3949 tasks. Only when there is no more work, a task will totally | |
3950 drain its local queue. | |
3951 | |
3952 (3) Global Mark Stack. This handles local queue overflow. During | |
3953 marking only sets of entries are moved between it and the local | |
3954 queues, as access to it requires a mutex and more fine-grain | |
3955 interaction with it which might cause contention. If it | |
3956 overflows, then the marking phase should restart and iterate | |
3957 over the bitmap to identify gray objects. Throughout the marking | |
3958 phase, tasks attempt to keep the global mark stack at a small | |
3959 length but not totally empty, so that entries are available for | |
3960 popping by other tasks. Only when there is no more work, tasks | |
3961 will totally drain the global mark stack. | |
3962 | |
3963 (4) Global Region Stack. Entries on it correspond to areas of | |
3964 the bitmap that need to be scanned since they contain gray | |
3965 objects. Pushes on the region stack only happen during | |
3966 evacuation pauses and typically correspond to areas covered by | |
3967 GC LABS. If it overflows, then the marking phase should restart | |
3968 and iterate over the bitmap to identify gray objects. Tasks will | |
3969 try to totally drain the region stack as soon as possible. | |
3970 | |
3971 (5) SATB Buffer Queue. This is where completed SATB buffers are | |
3972 made available. Buffers are regularly removed from this queue | |
3973 and scanned for roots, so that the queue doesn't get too | |
3974 long. During remark, all completed buffers are processed, as | |
3975 well as the filled in parts of any uncompleted buffers. | |
3976 | |
3977 The do_marking_step() method tries to abort when the time target | |
3978 has been reached. There are a few other cases when the | |
3979 do_marking_step() method also aborts: | |
3980 | |
3981 (1) When the marking phase has been aborted (after a Full GC). | |
3982 | |
3983 (2) When a global overflow (either on the global stack or the | |
3984 region stack) has been triggered. Before the task aborts, it | |
3985 will actually sync up with the other tasks to ensure that all | |
3986 the marking data structures (local queues, stacks, fingers etc.) | |
3987 are re-initialised so that when do_marking_step() completes, | |
3988 the marking phase can immediately restart. | |
3989 | |
3990 (3) When enough completed SATB buffers are available. The | |
3991 do_marking_step() method only tries to drain SATB buffers right | |
3992 at the beginning. So, if enough buffers are available, the | |
3993 marking step aborts and the SATB buffers are processed at | |
3994 the beginning of the next invocation. | |
3995 | |
3996 (4) To yield. when we have to yield then we abort and yield | |
3997 right at the end of do_marking_step(). This saves us from a lot | |
3998 of hassle as, by yielding we might allow a Full GC. If this | |
3999 happens then objects will be compacted underneath our feet, the | |
4000 heap might shrink, etc. We save checking for this by just | |
4001 aborting and doing the yield right at the end. | |
4002 | |
4003 From the above it follows that the do_marking_step() method should | |
4004 be called in a loop (or, otherwise, regularly) until it completes. | |
4005 | |
4006 If a marking step completes without its has_aborted() flag being | |
4007 true, it means it has completed the current marking phase (and | |
4008 also all other marking tasks have done so and have all synced up). | |
4009 | |
4010 A method called regular_clock_call() is invoked "regularly" (in | |
4011 sub ms intervals) throughout marking. It is this clock method that | |
4012 checks all the abort conditions which were mentioned above and | |
4013 decides when the task should abort. A work-based scheme is used to | |
4014 trigger this clock method: when the number of object words the | |
4015 marking phase has scanned or the number of references the marking | |
4016 phase has visited reach a given limit. Additional invocations to | |
4017 the method clock have been planted in a few other strategic places | |
4018 too. The initial reason for the clock method was to avoid calling | |
4019 vtime too regularly, as it is quite expensive. So, once it was in | |
4020 place, it was natural to piggy-back all the other conditions on it | |
4021 too and not constantly check them throughout the code. | |
4022 | |
4023 *****************************************************************************/ | |
4024 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4025 void CMTask::do_marking_step(double time_target_ms, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4026 bool do_stealing, |
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4027 bool do_termination) { |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4028 assert(time_target_ms >= 1.0, "minimum granularity is 1ms"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4029 assert(concurrent() == _cm->concurrent(), "they should be the same"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4030 |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4031 assert(concurrent() || _cm->region_stack_empty(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4032 "the region stack should have been cleared before remark"); |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
4033 assert(concurrent() || !_cm->has_aborted_regions(), |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
4034 "aborted regions should have been cleared before remark"); |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4035 assert(_region_finger == NULL, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4036 "this should be non-null only when a region is being scanned"); |
342 | 4037 |
4038 G1CollectorPolicy* g1_policy = _g1h->g1_policy(); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4039 assert(_task_queues != NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4040 assert(_task_queue != NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4041 assert(_task_queues->queue(_task_id) == _task_queue, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4042 |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4043 assert(!_claimed, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4044 "only one thread should claim this task at any one time"); |
342 | 4045 |
4046 // OK, this doesn't safeguard again all possible scenarios, as it is | |
4047 // possible for two threads to set the _claimed flag at the same | |
4048 // time. But it is only for debugging purposes anyway and it will | |
4049 // catch most problems. | |
4050 _claimed = true; | |
4051 | |
4052 _start_time_ms = os::elapsedVTime() * 1000.0; | |
4053 statsOnly( _interval_start_time_ms = _start_time_ms ); | |
4054 | |
4055 double diff_prediction_ms = | |
4056 g1_policy->get_new_prediction(&_marking_step_diffs_ms); | |
4057 _time_target_ms = time_target_ms - diff_prediction_ms; | |
4058 | |
4059 // set up the variables that are used in the work-based scheme to | |
4060 // call the regular clock method | |
4061 _words_scanned = 0; | |
4062 _refs_reached = 0; | |
4063 recalculate_limits(); | |
4064 | |
4065 // clear all flags | |
4066 clear_has_aborted(); | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4067 _has_timed_out = false; |
342 | 4068 _draining_satb_buffers = false; |
4069 | |
4070 ++_calls; | |
4071 | |
4072 if (_cm->verbose_low()) | |
4073 gclog_or_tty->print_cr("[%d] >>>>>>>>>> START, call = %d, " | |
4074 "target = %1.2lfms >>>>>>>>>>", | |
4075 _task_id, _calls, _time_target_ms); | |
4076 | |
4077 // Set up the bitmap and oop closures. Anything that uses them is | |
4078 // eventually called from this method, so it is OK to allocate these | |
4079 // statically. | |
4080 CMBitMapClosure bitmap_closure(this, _cm, _nextMarkBitMap); | |
4081 CMOopClosure oop_closure(_g1h, _cm, this); | |
4082 set_oop_closure(&oop_closure); | |
4083 | |
4084 if (_cm->has_overflown()) { | |
4085 // This can happen if the region stack or the mark stack overflows | |
4086 // during a GC pause and this task, after a yield point, | |
4087 // restarts. We have to abort as we need to get into the overflow | |
4088 // protocol which happens right at the end of this task. | |
4089 set_has_aborted(); | |
4090 } | |
4091 | |
4092 // First drain any available SATB buffers. After this, we will not | |
4093 // look at SATB buffers before the next invocation of this method. | |
4094 // If enough completed SATB buffers are queued up, the regular clock | |
4095 // will abort this task so that it restarts. | |
4096 drain_satb_buffers(); | |
4097 // ...then partially drain the local queue and the global stack | |
4098 drain_local_queue(true); | |
4099 drain_global_stack(true); | |
4100 | |
4101 // Then totally drain the region stack. We will not look at | |
4102 // it again before the next invocation of this method. Entries on | |
4103 // the region stack are only added during evacuation pauses, for | |
4104 // which we have to yield. When we do, we abort the task anyway so | |
4105 // it will look at the region stack again when it restarts. | |
4106 bitmap_closure.set_scanning_heap_region(false); | |
4107 drain_region_stack(&bitmap_closure); | |
4108 // ...then partially drain the local queue and the global stack | |
4109 drain_local_queue(true); | |
4110 drain_global_stack(true); | |
4111 | |
4112 do { | |
4113 if (!has_aborted() && _curr_region != NULL) { | |
4114 // This means that we're already holding on to a region. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4115 assert(_finger != NULL, "if region is not NULL, then the finger " |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4116 "should not be NULL either"); |
342 | 4117 |
4118 // We might have restarted this task after an evacuation pause | |
4119 // which might have evacuated the region we're holding on to | |
4120 // underneath our feet. Let's read its limit again to make sure | |
4121 // that we do not iterate over a region of the heap that | |
4122 // contains garbage (update_region_limit() will also move | |
4123 // _finger to the start of the region if it is found empty). | |
4124 update_region_limit(); | |
4125 // We will start from _finger not from the start of the region, | |
4126 // as we might be restarting this task after aborting half-way | |
4127 // through scanning this region. In this case, _finger points to | |
4128 // the address where we last found a marked object. If this is a | |
4129 // fresh region, _finger points to start(). | |
4130 MemRegion mr = MemRegion(_finger, _region_limit); | |
4131 | |
4132 if (_cm->verbose_low()) | |
4133 gclog_or_tty->print_cr("[%d] we're scanning part " | |
4134 "["PTR_FORMAT", "PTR_FORMAT") " | |
4135 "of region "PTR_FORMAT, | |
4136 _task_id, _finger, _region_limit, _curr_region); | |
4137 | |
4138 // Let's iterate over the bitmap of the part of the | |
4139 // region that is left. | |
4140 bitmap_closure.set_scanning_heap_region(true); | |
4141 if (mr.is_empty() || | |
4142 _nextMarkBitMap->iterate(&bitmap_closure, mr)) { | |
4143 // We successfully completed iterating over the region. Now, | |
4144 // let's give up the region. | |
4145 giveup_current_region(); | |
4146 regular_clock_call(); | |
4147 } else { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4148 assert(has_aborted(), "currently the only way to do so"); |
342 | 4149 // The only way to abort the bitmap iteration is to return |
4150 // false from the do_bit() method. However, inside the | |
4151 // do_bit() method we move the _finger to point to the | |
4152 // object currently being looked at. So, if we bail out, we | |
4153 // have definitely set _finger to something non-null. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4154 assert(_finger != NULL, "invariant"); |
342 | 4155 |
4156 // Region iteration was actually aborted. So now _finger | |
4157 // points to the address of the object we last scanned. If we | |
4158 // leave it there, when we restart this task, we will rescan | |
4159 // the object. It is easy to avoid this. We move the finger by | |
4160 // enough to point to the next possible object header (the | |
4161 // bitmap knows by how much we need to move it as it knows its | |
4162 // granularity). | |
1314
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4163 assert(_finger < _region_limit, "invariant"); |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4164 HeapWord* new_finger = _nextMarkBitMap->nextWord(_finger); |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4165 // Check if bitmap iteration was aborted while scanning the last object |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4166 if (new_finger >= _region_limit) { |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4167 giveup_current_region(); |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4168 } else { |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4169 move_finger_to(new_finger); |
3f0549ed0c98
6921710: G1: assert(new_finger >= _finger && new_finger < _region_limit,"invariant")
apetrusenko
parents:
1284
diff
changeset
|
4170 } |
342 | 4171 } |
4172 } | |
4173 // At this point we have either completed iterating over the | |
4174 // region we were holding on to, or we have aborted. | |
4175 | |
4176 // We then partially drain the local queue and the global stack. | |
4177 // (Do we really need this?) | |
4178 drain_local_queue(true); | |
4179 drain_global_stack(true); | |
4180 | |
4181 // Read the note on the claim_region() method on why it might | |
4182 // return NULL with potentially more regions available for | |
4183 // claiming and why we have to check out_of_regions() to determine | |
4184 // whether we're done or not. | |
4185 while (!has_aborted() && _curr_region == NULL && !_cm->out_of_regions()) { | |
4186 // We are going to try to claim a new region. We should have | |
4187 // given up on the previous one. | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4188 // Separated the asserts so that we know which one fires. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4189 assert(_curr_region == NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4190 assert(_finger == NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4191 assert(_region_limit == NULL, "invariant"); |
342 | 4192 if (_cm->verbose_low()) |
4193 gclog_or_tty->print_cr("[%d] trying to claim a new region", _task_id); | |
4194 HeapRegion* claimed_region = _cm->claim_region(_task_id); | |
4195 if (claimed_region != NULL) { | |
4196 // Yes, we managed to claim one | |
4197 statsOnly( ++_regions_claimed ); | |
4198 | |
4199 if (_cm->verbose_low()) | |
4200 gclog_or_tty->print_cr("[%d] we successfully claimed " | |
4201 "region "PTR_FORMAT, | |
4202 _task_id, claimed_region); | |
4203 | |
4204 setup_for_region(claimed_region); | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4205 assert(_curr_region == claimed_region, "invariant"); |
342 | 4206 } |
4207 // It is important to call the regular clock here. It might take | |
4208 // a while to claim a region if, for example, we hit a large | |
4209 // block of empty regions. So we need to call the regular clock | |
4210 // method once round the loop to make sure it's called | |
4211 // frequently enough. | |
4212 regular_clock_call(); | |
4213 } | |
4214 | |
4215 if (!has_aborted() && _curr_region == NULL) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4216 assert(_cm->out_of_regions(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4217 "at this point we should be out of regions"); |
342 | 4218 } |
4219 } while ( _curr_region != NULL && !has_aborted()); | |
4220 | |
4221 if (!has_aborted()) { | |
4222 // We cannot check whether the global stack is empty, since other | |
343
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4223 // tasks might be pushing objects to it concurrently. We also cannot |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4224 // check if the region stack is empty because if a thread is aborting |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4225 // it can push a partially done region back. |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4226 assert(_cm->out_of_regions(), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4227 "at this point we should be out of regions"); |
342 | 4228 |
4229 if (_cm->verbose_low()) | |
4230 gclog_or_tty->print_cr("[%d] all regions claimed", _task_id); | |
4231 | |
4232 // Try to reduce the number of available SATB buffers so that | |
4233 // remark has less work to do. | |
4234 drain_satb_buffers(); | |
4235 } | |
4236 | |
4237 // Since we've done everything else, we can now totally drain the | |
4238 // local queue and global stack. | |
4239 drain_local_queue(false); | |
4240 drain_global_stack(false); | |
4241 | |
4242 // Attempt at work stealing from other task's queues. | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4243 if (do_stealing && !has_aborted()) { |
342 | 4244 // We have not aborted. This means that we have finished all that |
4245 // we could. Let's try to do some stealing... | |
4246 | |
4247 // We cannot check whether the global stack is empty, since other | |
343
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4248 // tasks might be pushing objects to it concurrently. We also cannot |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4249 // check if the region stack is empty because if a thread is aborting |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4250 // it can push a partially done region back. |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4251 assert(_cm->out_of_regions() && _task_queue->size() == 0, |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4252 "only way to reach here"); |
342 | 4253 |
4254 if (_cm->verbose_low()) | |
4255 gclog_or_tty->print_cr("[%d] starting to steal", _task_id); | |
4256 | |
4257 while (!has_aborted()) { | |
4258 oop obj; | |
4259 statsOnly( ++_steal_attempts ); | |
4260 | |
4261 if (_cm->try_stealing(_task_id, &_hash_seed, obj)) { | |
4262 if (_cm->verbose_medium()) | |
4263 gclog_or_tty->print_cr("[%d] stolen "PTR_FORMAT" successfully", | |
4264 _task_id, (void*) obj); | |
4265 | |
4266 statsOnly( ++_steals ); | |
4267 | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4268 assert(_nextMarkBitMap->isMarked((HeapWord*) obj), |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4269 "any stolen object should be marked"); |
342 | 4270 scan_object(obj); |
4271 | |
4272 // And since we're towards the end, let's totally drain the | |
4273 // local queue and global stack. | |
4274 drain_local_queue(false); | |
4275 drain_global_stack(false); | |
4276 } else { | |
4277 break; | |
4278 } | |
4279 } | |
4280 } | |
4281 | |
4282 // We still haven't aborted. Now, let's try to get into the | |
4283 // termination protocol. | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4284 if (do_termination && !has_aborted()) { |
342 | 4285 // We cannot check whether the global stack is empty, since other |
343
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4286 // tasks might be concurrently pushing objects on it. We also cannot |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4287 // check if the region stack is empty because if a thread is aborting |
afc1ce1efe66
6710665: G1: guarantee(_cm->out_of_regions() && _cm->region_stack_empty() && _task_queue->size() == 0, ...)
iveresov
parents:
342
diff
changeset
|
4288 // it can push a partially done region back. |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4289 // Separated the asserts so that we know which one fires. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4290 assert(_cm->out_of_regions(), "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4291 assert(_task_queue->size() == 0, "only way to reach here"); |
342 | 4292 |
4293 if (_cm->verbose_low()) | |
4294 gclog_or_tty->print_cr("[%d] starting termination protocol", _task_id); | |
4295 | |
4296 _termination_start_time_ms = os::elapsedVTime() * 1000.0; | |
4297 // The CMTask class also extends the TerminatorTerminator class, | |
4298 // hence its should_exit_termination() method will also decide | |
4299 // whether to exit the termination protocol or not. | |
4300 bool finished = _cm->terminator()->offer_termination(this); | |
4301 double termination_end_time_ms = os::elapsedVTime() * 1000.0; | |
4302 _termination_time_ms += | |
4303 termination_end_time_ms - _termination_start_time_ms; | |
4304 | |
4305 if (finished) { | |
4306 // We're all done. | |
4307 | |
4308 if (_task_id == 0) { | |
4309 // let's allow task 0 to do this | |
4310 if (concurrent()) { | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4311 assert(_cm->concurrent_marking_in_progress(), "invariant"); |
342 | 4312 // we need to set this to false before the next |
4313 // safepoint. This way we ensure that the marking phase | |
4314 // doesn't observe any more heap expansions. | |
4315 _cm->clear_concurrent_marking_in_progress(); | |
4316 } | |
4317 } | |
4318 | |
4319 // We can now guarantee that the global stack is empty, since | |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4320 // all other tasks have finished. We separated the guarantees so |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4321 // that, if a condition is false, we can immediately find out |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4322 // which one. |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4323 guarantee(_cm->out_of_regions(), "only way to reach here"); |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
4324 guarantee(_aborted_region.is_empty(), "only way to reach here"); |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4325 guarantee(_cm->region_stack_empty(), "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4326 guarantee(_cm->mark_stack_empty(), "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4327 guarantee(_task_queue->size() == 0, "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4328 guarantee(!_cm->has_overflown(), "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4329 guarantee(!_cm->mark_stack_overflow(), "only way to reach here"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4330 guarantee(!_cm->region_stack_overflow(), "only way to reach here"); |
342 | 4331 |
4332 if (_cm->verbose_low()) | |
4333 gclog_or_tty->print_cr("[%d] all tasks terminated", _task_id); | |
4334 } else { | |
4335 // Apparently there's more work to do. Let's abort this task. It | |
4336 // will restart it and we can hopefully find more things to do. | |
4337 | |
4338 if (_cm->verbose_low()) | |
4339 gclog_or_tty->print_cr("[%d] apparently there is more work to do", _task_id); | |
4340 | |
4341 set_has_aborted(); | |
4342 statsOnly( ++_aborted_termination ); | |
4343 } | |
4344 } | |
4345 | |
4346 // Mainly for debugging purposes to make sure that a pointer to the | |
4347 // closure which was statically allocated in this frame doesn't | |
4348 // escape it by accident. | |
4349 set_oop_closure(NULL); | |
4350 double end_time_ms = os::elapsedVTime() * 1000.0; | |
4351 double elapsed_time_ms = end_time_ms - _start_time_ms; | |
4352 // Update the step history. | |
4353 _step_times_ms.add(elapsed_time_ms); | |
4354 | |
4355 if (has_aborted()) { | |
4356 // The task was aborted for some reason. | |
4357 | |
4358 statsOnly( ++_aborted ); | |
4359 | |
2174
234761c55641
6608385: G1: need to support parallel reference processing
johnc
parents:
2152
diff
changeset
|
4360 if (_has_timed_out) { |
342 | 4361 double diff_ms = elapsed_time_ms - _time_target_ms; |
4362 // Keep statistics of how well we did with respect to hitting | |
4363 // our target only if we actually timed out (if we aborted for | |
4364 // other reasons, then the results might get skewed). | |
4365 _marking_step_diffs_ms.add(diff_ms); | |
4366 } | |
4367 | |
4368 if (_cm->has_overflown()) { | |
4369 // This is the interesting one. We aborted because a global | |
4370 // overflow was raised. This means we have to restart the | |
4371 // marking phase and start iterating over regions. However, in | |
4372 // order to do this we have to make sure that all tasks stop | |
4373 // what they are doing and re-initialise in a safe manner. We | |
4374 // will achieve this with the use of two barrier sync points. | |
4375 | |
4376 if (_cm->verbose_low()) | |
4377 gclog_or_tty->print_cr("[%d] detected overflow", _task_id); | |
4378 | |
4379 _cm->enter_first_sync_barrier(_task_id); | |
4380 // When we exit this sync barrier we know that all tasks have | |
4381 // stopped doing marking work. So, it's now safe to | |
4382 // re-initialise our data structures. At the end of this method, | |
4383 // task 0 will clear the global data structures. | |
4384 | |
4385 statsOnly( ++_aborted_overflow ); | |
4386 | |
4387 // We clear the local state of this task... | |
4388 clear_region_fields(); | |
4389 | |
4390 // ...and enter the second barrier. | |
4391 _cm->enter_second_sync_barrier(_task_id); | |
4392 // At this point everything has bee re-initialised and we're | |
4393 // ready to restart. | |
4394 } | |
4395 | |
4396 if (_cm->verbose_low()) { | |
4397 gclog_or_tty->print_cr("[%d] <<<<<<<<<< ABORTING, target = %1.2lfms, " | |
4398 "elapsed = %1.2lfms <<<<<<<<<<", | |
4399 _task_id, _time_target_ms, elapsed_time_ms); | |
4400 if (_cm->has_aborted()) | |
4401 gclog_or_tty->print_cr("[%d] ========== MARKING ABORTED ==========", | |
4402 _task_id); | |
4403 } | |
4404 } else { | |
4405 if (_cm->verbose_low()) | |
4406 gclog_or_tty->print_cr("[%d] <<<<<<<<<< FINISHED, target = %1.2lfms, " | |
4407 "elapsed = %1.2lfms <<<<<<<<<<", | |
4408 _task_id, _time_target_ms, elapsed_time_ms); | |
4409 } | |
4410 | |
4411 _claimed = false; | |
4412 } | |
4413 | |
4414 CMTask::CMTask(int task_id, | |
4415 ConcurrentMark* cm, | |
4416 CMTaskQueue* task_queue, | |
4417 CMTaskQueueSet* task_queues) | |
4418 : _g1h(G1CollectedHeap::heap()), | |
4419 _task_id(task_id), _cm(cm), | |
4420 _claimed(false), | |
4421 _nextMarkBitMap(NULL), _hash_seed(17), | |
4422 _task_queue(task_queue), | |
4423 _task_queues(task_queues), | |
1835
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
4424 _oop_closure(NULL), |
4805b9f4779e
6941395: G1: Use only lock-free versions of region stack push() and pop()
johnc
parents:
1833
diff
changeset
|
4425 _aborted_region(MemRegion()) { |
1023
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4426 guarantee(task_queue != NULL, "invariant"); |
11d4857fe5e1
6888619: G1: too many guarantees in concurrent marking
tonyp
parents:
1022
diff
changeset
|
4427 guarantee(task_queues != NULL, "invariant"); |
342 | 4428 |
4429 statsOnly( _clock_due_to_scanning = 0; | |
4430 _clock_due_to_marking = 0 ); | |
4431 | |
4432 _marking_step_diffs_ms.add(0.5); | |
4433 } | |
2435
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4434 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4435 // These are formatting macros that are used below to ensure |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4436 // consistent formatting. The *_H_* versions are used to format the |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4437 // header for a particular value and they should be kept consistent |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4438 // with the corresponding macro. Also note that most of the macros add |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4439 // the necessary white space (as a prefix) which makes them a bit |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4440 // easier to compose. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4441 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4442 // All the output lines are prefixed with this string to be able to |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4443 // identify them easily in a large log file. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4444 #define G1PPRL_LINE_PREFIX "###" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4445 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4446 #define G1PPRL_ADDR_BASE_FORMAT " "PTR_FORMAT"-"PTR_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4447 #ifdef _LP64 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4448 #define G1PPRL_ADDR_BASE_H_FORMAT " %37s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4449 #else // _LP64 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4450 #define G1PPRL_ADDR_BASE_H_FORMAT " %21s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4451 #endif // _LP64 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4452 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4453 // For per-region info |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4454 #define G1PPRL_TYPE_FORMAT " %-4s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4455 #define G1PPRL_TYPE_H_FORMAT " %4s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4456 #define G1PPRL_BYTE_FORMAT " "SIZE_FORMAT_W(9) |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4457 #define G1PPRL_BYTE_H_FORMAT " %9s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4458 #define G1PPRL_DOUBLE_FORMAT " %14.1f" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4459 #define G1PPRL_DOUBLE_H_FORMAT " %14s" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4460 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4461 // For summary info |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4462 #define G1PPRL_SUM_ADDR_FORMAT(tag) " "tag":"G1PPRL_ADDR_BASE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4463 #define G1PPRL_SUM_BYTE_FORMAT(tag) " "tag": "SIZE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4464 #define G1PPRL_SUM_MB_FORMAT(tag) " "tag": %1.2f MB" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4465 #define G1PPRL_SUM_MB_PERC_FORMAT(tag) G1PPRL_SUM_MB_FORMAT(tag)" / %1.2f %%" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4466 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4467 G1PrintRegionLivenessInfoClosure:: |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4468 G1PrintRegionLivenessInfoClosure(outputStream* out, const char* phase_name) |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4469 : _out(out), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4470 _total_used_bytes(0), _total_capacity_bytes(0), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4471 _total_prev_live_bytes(0), _total_next_live_bytes(0), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4472 _hum_used_bytes(0), _hum_capacity_bytes(0), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4473 _hum_prev_live_bytes(0), _hum_next_live_bytes(0) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4474 G1CollectedHeap* g1h = G1CollectedHeap::heap(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4475 MemRegion g1_committed = g1h->g1_committed(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4476 MemRegion g1_reserved = g1h->g1_reserved(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4477 double now = os::elapsedTime(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4478 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4479 // Print the header of the output. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4480 _out->cr(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4481 _out->print_cr(G1PPRL_LINE_PREFIX" PHASE %s @ %1.3f", phase_name, now); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4482 _out->print_cr(G1PPRL_LINE_PREFIX" HEAP" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4483 G1PPRL_SUM_ADDR_FORMAT("committed") |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4484 G1PPRL_SUM_ADDR_FORMAT("reserved") |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4485 G1PPRL_SUM_BYTE_FORMAT("region-size"), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4486 g1_committed.start(), g1_committed.end(), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4487 g1_reserved.start(), g1_reserved.end(), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4488 HeapRegion::GrainBytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4489 _out->print_cr(G1PPRL_LINE_PREFIX); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4490 _out->print_cr(G1PPRL_LINE_PREFIX |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4491 G1PPRL_TYPE_H_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4492 G1PPRL_ADDR_BASE_H_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4493 G1PPRL_BYTE_H_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4494 G1PPRL_BYTE_H_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4495 G1PPRL_BYTE_H_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4496 G1PPRL_DOUBLE_H_FORMAT, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4497 "type", "address-range", |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4498 "used", "prev-live", "next-live", "gc-eff"); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4499 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4500 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4501 // It takes as a parameter a reference to one of the _hum_* fields, it |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4502 // deduces the corresponding value for a region in a humongous region |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4503 // series (either the region size, or what's left if the _hum_* field |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4504 // is < the region size), and updates the _hum_* field accordingly. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4505 size_t G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* hum_bytes) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4506 size_t bytes = 0; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4507 // The > 0 check is to deal with the prev and next live bytes which |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4508 // could be 0. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4509 if (*hum_bytes > 0) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4510 bytes = MIN2((size_t) HeapRegion::GrainBytes, *hum_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4511 *hum_bytes -= bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4512 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4513 return bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4514 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4515 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4516 // It deduces the values for a region in a humongous region series |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4517 // from the _hum_* fields and updates those accordingly. It assumes |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4518 // that that _hum_* fields have already been set up from the "starts |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4519 // humongous" region and we visit the regions in address order. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4520 void G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* used_bytes, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4521 size_t* capacity_bytes, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4522 size_t* prev_live_bytes, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4523 size_t* next_live_bytes) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4524 assert(_hum_used_bytes > 0 && _hum_capacity_bytes > 0, "pre-condition"); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4525 *used_bytes = get_hum_bytes(&_hum_used_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4526 *capacity_bytes = get_hum_bytes(&_hum_capacity_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4527 *prev_live_bytes = get_hum_bytes(&_hum_prev_live_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4528 *next_live_bytes = get_hum_bytes(&_hum_next_live_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4529 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4530 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4531 bool G1PrintRegionLivenessInfoClosure::doHeapRegion(HeapRegion* r) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4532 const char* type = ""; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4533 HeapWord* bottom = r->bottom(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4534 HeapWord* end = r->end(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4535 size_t capacity_bytes = r->capacity(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4536 size_t used_bytes = r->used(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4537 size_t prev_live_bytes = r->live_bytes(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4538 size_t next_live_bytes = r->next_live_bytes(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4539 double gc_eff = r->gc_efficiency(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4540 if (r->used() == 0) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4541 type = "FREE"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4542 } else if (r->is_survivor()) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4543 type = "SURV"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4544 } else if (r->is_young()) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4545 type = "EDEN"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4546 } else if (r->startsHumongous()) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4547 type = "HUMS"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4548 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4549 assert(_hum_used_bytes == 0 && _hum_capacity_bytes == 0 && |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4550 _hum_prev_live_bytes == 0 && _hum_next_live_bytes == 0, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4551 "they should have been zeroed after the last time we used them"); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4552 // Set up the _hum_* fields. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4553 _hum_capacity_bytes = capacity_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4554 _hum_used_bytes = used_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4555 _hum_prev_live_bytes = prev_live_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4556 _hum_next_live_bytes = next_live_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4557 get_hum_bytes(&used_bytes, &capacity_bytes, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4558 &prev_live_bytes, &next_live_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4559 end = bottom + HeapRegion::GrainWords; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4560 } else if (r->continuesHumongous()) { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4561 type = "HUMC"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4562 get_hum_bytes(&used_bytes, &capacity_bytes, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4563 &prev_live_bytes, &next_live_bytes); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4564 assert(end == bottom + HeapRegion::GrainWords, "invariant"); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4565 } else { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4566 type = "OLD"; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4567 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4568 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4569 _total_used_bytes += used_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4570 _total_capacity_bytes += capacity_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4571 _total_prev_live_bytes += prev_live_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4572 _total_next_live_bytes += next_live_bytes; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4573 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4574 // Print a line for this particular region. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4575 _out->print_cr(G1PPRL_LINE_PREFIX |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4576 G1PPRL_TYPE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4577 G1PPRL_ADDR_BASE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4578 G1PPRL_BYTE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4579 G1PPRL_BYTE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4580 G1PPRL_BYTE_FORMAT |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4581 G1PPRL_DOUBLE_FORMAT, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4582 type, bottom, end, |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4583 used_bytes, prev_live_bytes, next_live_bytes, gc_eff); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4584 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4585 return false; |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4586 } |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4587 |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4588 G1PrintRegionLivenessInfoClosure::~G1PrintRegionLivenessInfoClosure() { |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4589 // Print the footer of the output. |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4590 _out->print_cr(G1PPRL_LINE_PREFIX); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4591 _out->print_cr(G1PPRL_LINE_PREFIX |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4592 " SUMMARY" |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4593 G1PPRL_SUM_MB_FORMAT("capacity") |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4594 G1PPRL_SUM_MB_PERC_FORMAT("used") |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4595 G1PPRL_SUM_MB_PERC_FORMAT("prev-live") |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4596 G1PPRL_SUM_MB_PERC_FORMAT("next-live"), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4597 bytes_to_mb(_total_capacity_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4598 bytes_to_mb(_total_used_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4599 perc(_total_used_bytes, _total_capacity_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4600 bytes_to_mb(_total_prev_live_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4601 perc(_total_prev_live_bytes, _total_capacity_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4602 bytes_to_mb(_total_next_live_bytes), |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4603 perc(_total_next_live_bytes, _total_capacity_bytes)); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4604 _out->cr(); |
371bbc844bf1
7027766: G1: introduce flag to dump the liveness information per region at the end of marking
tonyp
parents:
2369
diff
changeset
|
4605 } |