|
1 /* |
|
2 * Copyright (c) 2001, 2017, Oracle and/or its affiliates. All rights reserved. |
|
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
|
4 * |
|
5 * This code is free software; you can redistribute it and/or modify it |
|
6 * under the terms of the GNU General Public License version 2 only, as |
|
7 * published by the Free Software Foundation. |
|
8 * |
|
9 * This code is distributed in the hope that it will be useful, but WITHOUT |
|
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
|
12 * version 2 for more details (a copy is included in the LICENSE file that |
|
13 * accompanied this code). |
|
14 * |
|
15 * You should have received a copy of the GNU General Public License version |
|
16 * 2 along with this work; if not, write to the Free Software Foundation, |
|
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
|
18 * |
|
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
|
20 * or visit www.oracle.com if you need additional information or have any |
|
21 * questions. |
|
22 * |
|
23 */ |
|
24 |
|
25 #include "precompiled.hpp" |
|
26 #include "gc/g1/dirtyCardQueue.hpp" |
|
27 #include "gc/g1/g1CollectedHeap.inline.hpp" |
|
28 #include "gc/g1/g1RemSet.hpp" |
|
29 #include "gc/g1/heapRegionRemSet.hpp" |
|
30 #include "gc/shared/workgroup.hpp" |
|
31 #include "runtime/atomic.hpp" |
|
32 #include "runtime/mutexLocker.hpp" |
|
33 #include "runtime/safepoint.hpp" |
|
34 #include "runtime/thread.inline.hpp" |
|
35 |
|
36 // Closure used for updating remembered sets and recording references that |
|
37 // point into the collection set while the mutator is running. |
|
38 // Assumed to be only executed concurrently with the mutator. Yields via |
|
39 // SuspendibleThreadSet after every card. |
|
40 class G1RefineCardConcurrentlyClosure: public CardTableEntryClosure { |
|
41 public: |
|
42 bool do_card_ptr(jbyte* card_ptr, uint worker_i) { |
|
43 G1CollectedHeap::heap()->g1_rem_set()->refine_card_concurrently(card_ptr, worker_i); |
|
44 |
|
45 if (SuspendibleThreadSet::should_yield()) { |
|
46 // Caller will actually yield. |
|
47 return false; |
|
48 } |
|
49 // Otherwise, we finished successfully; return true. |
|
50 return true; |
|
51 } |
|
52 }; |
|
53 |
|
54 // Represents a set of free small integer ids. |
|
55 class FreeIdSet : public CHeapObj<mtGC> { |
|
56 enum { |
|
57 end_of_list = UINT_MAX, |
|
58 claimed = UINT_MAX - 1 |
|
59 }; |
|
60 |
|
61 uint _size; |
|
62 Monitor* _mon; |
|
63 |
|
64 uint* _ids; |
|
65 uint _hd; |
|
66 uint _waiters; |
|
67 uint _claimed; |
|
68 |
|
69 public: |
|
70 FreeIdSet(uint size, Monitor* mon); |
|
71 ~FreeIdSet(); |
|
72 |
|
73 // Returns an unclaimed parallel id (waiting for one to be released if |
|
74 // necessary). |
|
75 uint claim_par_id(); |
|
76 |
|
77 void release_par_id(uint id); |
|
78 }; |
|
79 |
|
80 FreeIdSet::FreeIdSet(uint size, Monitor* mon) : |
|
81 _size(size), _mon(mon), _hd(0), _waiters(0), _claimed(0) |
|
82 { |
|
83 guarantee(size != 0, "must be"); |
|
84 _ids = NEW_C_HEAP_ARRAY(uint, size, mtGC); |
|
85 for (uint i = 0; i < size - 1; i++) { |
|
86 _ids[i] = i+1; |
|
87 } |
|
88 _ids[size-1] = end_of_list; // end of list. |
|
89 } |
|
90 |
|
91 FreeIdSet::~FreeIdSet() { |
|
92 FREE_C_HEAP_ARRAY(uint, _ids); |
|
93 } |
|
94 |
|
95 uint FreeIdSet::claim_par_id() { |
|
96 MutexLockerEx x(_mon, Mutex::_no_safepoint_check_flag); |
|
97 while (_hd == end_of_list) { |
|
98 _waiters++; |
|
99 _mon->wait(Mutex::_no_safepoint_check_flag); |
|
100 _waiters--; |
|
101 } |
|
102 uint res = _hd; |
|
103 _hd = _ids[res]; |
|
104 _ids[res] = claimed; // For debugging. |
|
105 _claimed++; |
|
106 return res; |
|
107 } |
|
108 |
|
109 void FreeIdSet::release_par_id(uint id) { |
|
110 MutexLockerEx x(_mon, Mutex::_no_safepoint_check_flag); |
|
111 assert(_ids[id] == claimed, "Precondition."); |
|
112 _ids[id] = _hd; |
|
113 _hd = id; |
|
114 _claimed--; |
|
115 if (_waiters > 0) { |
|
116 _mon->notify_all(); |
|
117 } |
|
118 } |
|
119 |
|
120 DirtyCardQueue::DirtyCardQueue(DirtyCardQueueSet* qset, bool permanent) : |
|
121 // Dirty card queues are always active, so we create them with their |
|
122 // active field set to true. |
|
123 PtrQueue(qset, permanent, true /* active */) |
|
124 { } |
|
125 |
|
126 DirtyCardQueue::~DirtyCardQueue() { |
|
127 if (!is_permanent()) { |
|
128 flush(); |
|
129 } |
|
130 } |
|
131 |
|
132 DirtyCardQueueSet::DirtyCardQueueSet(bool notify_when_complete) : |
|
133 PtrQueueSet(notify_when_complete), |
|
134 _shared_dirty_card_queue(this, true /* permanent */), |
|
135 _free_ids(NULL), |
|
136 _processed_buffers_mut(0), _processed_buffers_rs_thread(0) |
|
137 { |
|
138 _all_active = true; |
|
139 } |
|
140 |
|
141 // Determines how many mutator threads can process the buffers in parallel. |
|
142 uint DirtyCardQueueSet::num_par_ids() { |
|
143 return (uint)os::initial_active_processor_count(); |
|
144 } |
|
145 |
|
146 void DirtyCardQueueSet::initialize(Monitor* cbl_mon, |
|
147 Mutex* fl_lock, |
|
148 int process_completed_threshold, |
|
149 int max_completed_queue, |
|
150 Mutex* lock, |
|
151 DirtyCardQueueSet* fl_owner, |
|
152 bool init_free_ids) { |
|
153 PtrQueueSet::initialize(cbl_mon, |
|
154 fl_lock, |
|
155 process_completed_threshold, |
|
156 max_completed_queue, |
|
157 fl_owner); |
|
158 set_buffer_size(G1UpdateBufferSize); |
|
159 _shared_dirty_card_queue.set_lock(lock); |
|
160 if (init_free_ids) { |
|
161 _free_ids = new FreeIdSet(num_par_ids(), _cbl_mon); |
|
162 } |
|
163 } |
|
164 |
|
165 void DirtyCardQueueSet::handle_zero_index_for_thread(JavaThread* t) { |
|
166 t->dirty_card_queue().handle_zero_index(); |
|
167 } |
|
168 |
|
169 bool DirtyCardQueueSet::apply_closure_to_buffer(CardTableEntryClosure* cl, |
|
170 BufferNode* node, |
|
171 bool consume, |
|
172 uint worker_i) { |
|
173 if (cl == NULL) return true; |
|
174 bool result = true; |
|
175 void** buf = BufferNode::make_buffer_from_node(node); |
|
176 size_t i = node->index(); |
|
177 size_t limit = buffer_size(); |
|
178 for ( ; i < limit; ++i) { |
|
179 jbyte* card_ptr = static_cast<jbyte*>(buf[i]); |
|
180 assert(card_ptr != NULL, "invariant"); |
|
181 if (!cl->do_card_ptr(card_ptr, worker_i)) { |
|
182 result = false; // Incomplete processing. |
|
183 break; |
|
184 } |
|
185 } |
|
186 if (consume) { |
|
187 assert(i <= buffer_size(), "invariant"); |
|
188 node->set_index(i); |
|
189 } |
|
190 return result; |
|
191 } |
|
192 |
|
193 #ifndef ASSERT |
|
194 #define assert_fully_consumed(node, buffer_size) |
|
195 #else |
|
196 #define assert_fully_consumed(node, buffer_size) \ |
|
197 do { \ |
|
198 size_t _afc_index = (node)->index(); \ |
|
199 size_t _afc_size = (buffer_size); \ |
|
200 assert(_afc_index == _afc_size, \ |
|
201 "Buffer was not fully consumed as claimed: index: " \ |
|
202 SIZE_FORMAT ", size: " SIZE_FORMAT, \ |
|
203 _afc_index, _afc_size); \ |
|
204 } while (0) |
|
205 #endif // ASSERT |
|
206 |
|
207 bool DirtyCardQueueSet::mut_process_buffer(BufferNode* node) { |
|
208 guarantee(_free_ids != NULL, "must be"); |
|
209 |
|
210 uint worker_i = _free_ids->claim_par_id(); // temporarily claim an id |
|
211 G1RefineCardConcurrentlyClosure cl; |
|
212 bool result = apply_closure_to_buffer(&cl, node, true, worker_i); |
|
213 _free_ids->release_par_id(worker_i); // release the id |
|
214 |
|
215 if (result) { |
|
216 assert_fully_consumed(node, buffer_size()); |
|
217 Atomic::inc(&_processed_buffers_mut); |
|
218 } |
|
219 return result; |
|
220 } |
|
221 |
|
222 |
|
223 BufferNode* DirtyCardQueueSet::get_completed_buffer(size_t stop_at) { |
|
224 BufferNode* nd = NULL; |
|
225 MutexLockerEx x(_cbl_mon, Mutex::_no_safepoint_check_flag); |
|
226 |
|
227 if (_n_completed_buffers <= stop_at) { |
|
228 _process_completed = false; |
|
229 return NULL; |
|
230 } |
|
231 |
|
232 if (_completed_buffers_head != NULL) { |
|
233 nd = _completed_buffers_head; |
|
234 assert(_n_completed_buffers > 0, "Invariant"); |
|
235 _completed_buffers_head = nd->next(); |
|
236 _n_completed_buffers--; |
|
237 if (_completed_buffers_head == NULL) { |
|
238 assert(_n_completed_buffers == 0, "Invariant"); |
|
239 _completed_buffers_tail = NULL; |
|
240 } |
|
241 } |
|
242 DEBUG_ONLY(assert_completed_buffer_list_len_correct_locked()); |
|
243 return nd; |
|
244 } |
|
245 |
|
246 bool DirtyCardQueueSet::refine_completed_buffer_concurrently(uint worker_i, size_t stop_at) { |
|
247 G1RefineCardConcurrentlyClosure cl; |
|
248 return apply_closure_to_completed_buffer(&cl, worker_i, stop_at, false); |
|
249 } |
|
250 |
|
251 bool DirtyCardQueueSet::apply_closure_during_gc(CardTableEntryClosure* cl, uint worker_i) { |
|
252 assert_at_safepoint(false); |
|
253 return apply_closure_to_completed_buffer(cl, worker_i, 0, true); |
|
254 } |
|
255 |
|
256 bool DirtyCardQueueSet::apply_closure_to_completed_buffer(CardTableEntryClosure* cl, |
|
257 uint worker_i, |
|
258 size_t stop_at, |
|
259 bool during_pause) { |
|
260 assert(!during_pause || stop_at == 0, "Should not leave any completed buffers during a pause"); |
|
261 BufferNode* nd = get_completed_buffer(stop_at); |
|
262 if (nd == NULL) { |
|
263 return false; |
|
264 } else { |
|
265 if (apply_closure_to_buffer(cl, nd, true, worker_i)) { |
|
266 assert_fully_consumed(nd, buffer_size()); |
|
267 // Done with fully processed buffer. |
|
268 deallocate_buffer(nd); |
|
269 Atomic::inc(&_processed_buffers_rs_thread); |
|
270 } else { |
|
271 // Return partially processed buffer to the queue. |
|
272 guarantee(!during_pause, "Should never stop early"); |
|
273 enqueue_complete_buffer(nd); |
|
274 } |
|
275 return true; |
|
276 } |
|
277 } |
|
278 |
|
279 void DirtyCardQueueSet::par_apply_closure_to_all_completed_buffers(CardTableEntryClosure* cl) { |
|
280 BufferNode* nd = _cur_par_buffer_node; |
|
281 while (nd != NULL) { |
|
282 BufferNode* next = nd->next(); |
|
283 void* actual = Atomic::cmpxchg_ptr(next, &_cur_par_buffer_node, nd); |
|
284 if (actual == nd) { |
|
285 bool b = apply_closure_to_buffer(cl, nd, false); |
|
286 guarantee(b, "Should not stop early."); |
|
287 nd = next; |
|
288 } else { |
|
289 nd = static_cast<BufferNode*>(actual); |
|
290 } |
|
291 } |
|
292 } |
|
293 |
|
294 // Deallocates any completed log buffers |
|
295 void DirtyCardQueueSet::clear() { |
|
296 BufferNode* buffers_to_delete = NULL; |
|
297 { |
|
298 MutexLockerEx x(_cbl_mon, Mutex::_no_safepoint_check_flag); |
|
299 while (_completed_buffers_head != NULL) { |
|
300 BufferNode* nd = _completed_buffers_head; |
|
301 _completed_buffers_head = nd->next(); |
|
302 nd->set_next(buffers_to_delete); |
|
303 buffers_to_delete = nd; |
|
304 } |
|
305 _n_completed_buffers = 0; |
|
306 _completed_buffers_tail = NULL; |
|
307 DEBUG_ONLY(assert_completed_buffer_list_len_correct_locked()); |
|
308 } |
|
309 while (buffers_to_delete != NULL) { |
|
310 BufferNode* nd = buffers_to_delete; |
|
311 buffers_to_delete = nd->next(); |
|
312 deallocate_buffer(nd); |
|
313 } |
|
314 |
|
315 } |
|
316 |
|
317 void DirtyCardQueueSet::abandon_logs() { |
|
318 assert(SafepointSynchronize::is_at_safepoint(), "Must be at safepoint."); |
|
319 clear(); |
|
320 // Since abandon is done only at safepoints, we can safely manipulate |
|
321 // these queues. |
|
322 for (JavaThread* t = Threads::first(); t; t = t->next()) { |
|
323 t->dirty_card_queue().reset(); |
|
324 } |
|
325 shared_dirty_card_queue()->reset(); |
|
326 } |
|
327 |
|
328 void DirtyCardQueueSet::concatenate_log(DirtyCardQueue& dcq) { |
|
329 if (!dcq.is_empty()) { |
|
330 dcq.flush(); |
|
331 } |
|
332 } |
|
333 |
|
334 void DirtyCardQueueSet::concatenate_logs() { |
|
335 // Iterate over all the threads, if we find a partial log add it to |
|
336 // the global list of logs. Temporarily turn off the limit on the number |
|
337 // of outstanding buffers. |
|
338 int save_max_completed_queue = _max_completed_queue; |
|
339 _max_completed_queue = max_jint; |
|
340 assert(SafepointSynchronize::is_at_safepoint(), "Must be at safepoint."); |
|
341 for (JavaThread* t = Threads::first(); t; t = t->next()) { |
|
342 concatenate_log(t->dirty_card_queue()); |
|
343 } |
|
344 concatenate_log(_shared_dirty_card_queue); |
|
345 // Restore the completed buffer queue limit. |
|
346 _max_completed_queue = save_max_completed_queue; |
|
347 } |