105 |
|
* successful atomic operation per enq/deq pair. But it also |
106 |
|
* enables lower cost variants of queue maintenance mechanics. (A |
107 |
|
* variation of this idea applies even for non-dual queues that |
108 |
< |
* support deletion of embedded elements, such as |
108 |
> |
* support deletion of interior elements, such as |
109 |
|
* j.u.c.ConcurrentLinkedQueue.) |
110 |
|
* |
111 |
< |
* Once a node is matched, its item can never again change. We |
112 |
< |
* may thus arrange that the linked list of them contains a prefix |
113 |
< |
* of zero or more matched nodes, followed by a suffix of zero or |
114 |
< |
* more unmatched nodes. (Note that we allow both the prefix and |
115 |
< |
* suffix to be zero length, which in turn means that we do not |
116 |
< |
* use a dummy header.) If we were not concerned with either time |
117 |
< |
* or space efficiency, we could correctly perform enqueue and |
118 |
< |
* dequeue operations by traversing from a pointer to the initial |
119 |
< |
* node; CASing the item of the first unmatched node on match and |
120 |
< |
* CASing the next field of the trailing node on appends. While |
121 |
< |
* this would be a terrible idea in itself, it does have the |
122 |
< |
* benefit of not requiring ANY atomic updates on head/tail |
123 |
< |
* fields. |
111 |
> |
* Once a node is matched, its match status can never again |
112 |
> |
* change. We may thus arrange that the linked list of them |
113 |
> |
* contain a prefix of zero or more matched nodes, followed by a |
114 |
> |
* suffix of zero or more unmatched nodes. (Note that we allow |
115 |
> |
* both the prefix and suffix to be zero length, which in turn |
116 |
> |
* means that we do not use a dummy header.) If we were not |
117 |
> |
* concerned with either time or space efficiency, we could |
118 |
> |
* correctly perform enqueue and dequeue operations by traversing |
119 |
> |
* from a pointer to the initial node; CASing the item of the |
120 |
> |
* first unmatched node on match and CASing the next field of the |
121 |
> |
* trailing node on appends. (Plus some special-casing when |
122 |
> |
* initially empty). While this would be a terrible idea in |
123 |
> |
* itself, it does have the benefit of not requiring ANY atomic |
124 |
> |
* updates on head/tail fields. |
125 |
|
* |
126 |
|
* We introduce here an approach that lies between the extremes of |
127 |
< |
* never versus always updating queue (head and tail) pointers |
128 |
< |
* that reflects the tradeoff of sometimes require extra traversal |
129 |
< |
* steps to locate the first and/or last unmatched nodes, versus |
130 |
< |
* the reduced overhead and contention of fewer updates to queue |
131 |
< |
* pointers. For example, a possible snapshot of a queue is: |
127 |
> |
* never versus always updating queue (head and tail) pointers. |
128 |
> |
* This offers a tradeoff between sometimes requiring extra |
129 |
> |
* traversal steps to locate the first and/or last unmatched |
130 |
> |
* nodes, versus the reduced overhead and contention of fewer |
131 |
> |
* updates to queue pointers. For example, a possible snapshot of |
132 |
> |
* a queue is: |
133 |
|
* |
134 |
|
* head tail |
135 |
|
* | | |
141 |
|
* similarly for "tail") is an empirical matter. We have found |
142 |
|
* that using very small constants in the range of 1-3 work best |
143 |
|
* over a range of platforms. Larger values introduce increasing |
144 |
< |
* costs of cache misses and risks of long traversal chains. |
144 |
> |
* costs of cache misses and risks of long traversal chains, while |
145 |
> |
* smaller values increase CAS contention and overhead. |
146 |
|
* |
147 |
|
* Dual queues with slack differ from plain M&S dual queues by |
148 |
|
* virtue of only sometimes updating head or tail pointers when |
161 |
|
* targets. Even when using very small slack values, this |
162 |
|
* approach works well for dual queues because it allows all |
163 |
|
* operations up to the point of matching or appending an item |
164 |
< |
* (hence potentially releasing another thread) to be read-only, |
165 |
< |
* thus not introducing any further contention. As described |
166 |
< |
* below, we implement this by performing slack maintenance |
167 |
< |
* retries only after these points. |
164 |
> |
* (hence potentially allowing progress by another thread) to be |
165 |
> |
* read-only, thus not introducing any further contention. As |
166 |
> |
* described below, we implement this by performing slack |
167 |
> |
* maintenance retries only after these points. |
168 |
|
* |
169 |
|
* As an accompaniment to such techniques, traversal overhead can |
170 |
|
* be further reduced without increasing contention of head |
171 |
< |
* pointer updates. During traversals, threads may sometimes |
172 |
< |
* shortcut the "next" link path from the current "head" node to |
173 |
< |
* be closer to the currently known first unmatched node. Again, |
174 |
< |
* this may be triggered with using thresholds or randomization. |
171 |
> |
* pointer updates: Threads may sometimes shortcut the "next" link |
172 |
> |
* path from the current "head" node to be closer to the currently |
173 |
> |
* known first unmatched node, and similarly for tail. Again, this |
174 |
> |
* may be triggered with using thresholds or randomization. |
175 |
|
* |
176 |
|
* These ideas must be further extended to avoid unbounded amounts |
177 |
|
* of costly-to-reclaim garbage caused by the sequential "next" |
183 |
|
* (Similar issues arise in non-GC environments.) To cope with |
184 |
|
* this in our implementation, upon CASing to advance the head |
185 |
|
* pointer, we set the "next" link of the previous head to point |
186 |
< |
* only to itself; thus limiting the length connected dead lists. |
186 |
> |
* only to itself; thus limiting the length of connected dead lists. |
187 |
|
* (We also take similar care to wipe out possibly garbage |
188 |
|
* retaining values held in other Node fields.) However, doing so |
189 |
|
* adds some further complexity to traversal: If any "next" |
199 |
|
* mechanics because an update may leave head at a detached node. |
200 |
|
* And while direct writes are possible for tail updates, they |
201 |
|
* increase the risk of long retraversals, and hence long garbage |
202 |
< |
* chains which can be much more costly than is worthwhile |
202 |
> |
* chains, which can be much more costly than is worthwhile |
203 |
|
* considering that the cost difference of performing a CAS vs |
204 |
|
* write is smaller when they are not triggered on each operation |
205 |
|
* (especially considering that writes and CASes equally require |
206 |
|
* additional GC bookkeeping ("write barriers") that are sometimes |
207 |
|
* more costly than the writes themselves because of contention). |
208 |
|
* |
209 |
< |
* Removal of internal nodes (due to timed out or interrupted |
210 |
< |
* waits, or calls to remove or Iterator.remove) uses a scheme |
211 |
< |
* roughly similar to that in Scherer, Lea, and Scott |
212 |
< |
* SynchronousQueue. Given a predecessor, we can unsplice any node |
213 |
< |
* except the (actual) tail of the queue. To avoid build-up of |
214 |
< |
* cancelled trailing nodes, upon a request to remove a trailing |
215 |
< |
* node, it is placed in field "cleanMe" to be unspliced later. |
209 |
> |
* Removal of interior nodes (due to timed out or interrupted |
210 |
> |
* waits, or calls to remove(x) or Iterator.remove) can use a |
211 |
> |
* scheme roughly similar to that described in Scherer, Lea, and |
212 |
> |
* Scott's SynchronousQueue. Given a predecessor, we can unsplice |
213 |
> |
* any node except the (actual) tail of the queue. To avoid |
214 |
> |
* build-up of cancelled trailing nodes, upon a request to remove |
215 |
> |
* a trailing node, it is placed in field "cleanMe" to be |
216 |
> |
* unspliced upon the next call to unsplice any other node. |
217 |
> |
* Situations needing such mechanics are not common but do occur |
218 |
> |
* in practice; for example when an unbounded series of short |
219 |
> |
* timed calls to poll repeatedly time out but never otherwise |
220 |
> |
* fall off the list because of an untimed call to take at the |
221 |
> |
* front of the queue. Note that maintaining field cleanMe does |
222 |
> |
* not otherwise much impact garbage retention even if never |
223 |
> |
* cleared by some other call because the held node will |
224 |
> |
* eventually either directly or indirectly lead to a self-link |
225 |
> |
* once off the list. |
226 |
|
* |
227 |
|
* *** Overview of implementation *** |
228 |
|
* |
229 |
< |
* We use a threshold-based approach to updates, with a target |
230 |
< |
* slack of two. The slack value is hard-wired: a path greater |
229 |
> |
* We use a threshold-based approach to updates, with a slack |
230 |
> |
* threshold of two -- that is, we update head/tail when the |
231 |
> |
* current pointer appears to be two or more steps away from the |
232 |
> |
* first/last node. The slack value is hard-wired: a path greater |
233 |
|
* than one is naturally implemented by checking equality of |
234 |
|
* traversal pointers except when the list has only one element, |
235 |
< |
* in which case we keep max slack at one. Avoiding tracking |
236 |
< |
* explicit counts across situations slightly simplifies an |
235 |
> |
* in which case we keep slack threshold at one. Avoiding tracking |
236 |
> |
* explicit counts across method calls slightly simplifies an |
237 |
|
* already-messy implementation. Using randomization would |
238 |
|
* probably work better if there were a low-quality dirt-cheap |
239 |
|
* per-thread one available, but even ThreadLocalRandom is too |
240 |
|
* heavy for these purposes. |
241 |
|
* |
242 |
< |
* With such a small slack value, path short-circuiting is rarely |
243 |
< |
* worthwhile. However, it is used (in awaitMatch) immediately |
244 |
< |
* before a waiting thread starts to block, as a final bit of |
245 |
< |
* helping at a point when contention with others is extremely |
246 |
< |
* unlikely (since if other threads that could release it are |
247 |
< |
* operating, then the current thread wouldn't be blocking). |
242 |
> |
* With such a small slack threshold value, it is rarely |
243 |
> |
* worthwhile to augment this with path short-circuiting; i.e., |
244 |
> |
* unsplicing nodes between head and the first unmatched node, or |
245 |
> |
* similarly for tail, rather than advancing head or tail |
246 |
> |
* proper. However, it is used (in awaitMatch) immediately before |
247 |
> |
* a waiting thread starts to block, as a final bit of helping at |
248 |
> |
* a point when contention with others is extremely unlikely |
249 |
> |
* (since if other threads that could release it are operating, |
250 |
> |
* then the current thread wouldn't be blocking). |
251 |
> |
* |
252 |
> |
* We allow both the head and tail fields to be null before any |
253 |
> |
* nodes are enqueued; initializing upon first append. This |
254 |
> |
* simplifies some other logic, as well as providing more |
255 |
> |
* efficient explicit control paths instead of letting JVMs insert |
256 |
> |
* implicit NullPointerExceptions when they are null. While not |
257 |
> |
* currently fully implemented, we also leave open the possibility |
258 |
> |
* of re-nulling these fields when empty (which is complicated to |
259 |
> |
* arrange, for little benefit.) |
260 |
|
* |
261 |
|
* All enqueue/dequeue operations are handled by the single method |
262 |
|
* "xfer" with parameters indicating whether to act as some form |
263 |
|
* of offer, put, poll, take, or transfer (each possibly with |
264 |
|
* timeout). The relative complexity of using one monolithic |
265 |
|
* method outweighs the code bulk and maintenance problems of |
266 |
< |
* using nine separate methods. |
266 |
> |
* using separate methods for each case. |
267 |
|
* |
268 |
|
* Operation consists of up to three phases. The first is |
269 |
|
* implemented within method xfer, the second in tryAppend, and |
276 |
|
* case matching it and returning, also if necessary updating |
277 |
|
* head to one past the matched node (or the node itself if the |
278 |
|
* list has no other unmatched nodes). If the CAS misses, then |
279 |
< |
* a retry loops until the slack is at most two. Traversals |
280 |
< |
* also check if the initial head is now off-list, in which |
281 |
< |
* case they start at the new head. |
279 |
> |
* a loop retries advancing head by two steps until either |
280 |
> |
* success or the slack is at most two. By requiring that each |
281 |
> |
* attempt advances head by two (if applicable), we ensure that |
282 |
> |
* the slack does not grow without bound. Traversals also check |
283 |
> |
* if the initial head is now off-list, in which case they |
284 |
> |
* start at the new head. |
285 |
|
* |
286 |
|
* If no candidates are found and the call was untimed |
287 |
|
* poll/offer, (argument "how" is NOW) return. |
288 |
|
* |
289 |
|
* 2. Try to append a new node (method tryAppend) |
290 |
|
* |
291 |
< |
* Starting at current tail pointer, try to append a new node |
292 |
< |
* to the list (or if head was null, establish the first |
293 |
< |
* node). Nodes can be appended only if their predecessors are |
294 |
< |
* either already matched or are of the same mode. If we detect |
295 |
< |
* otherwise, then a new node with opposite mode must have been |
296 |
< |
* appended during traversal, so must restart at phase 1. The |
297 |
< |
* traversal and update steps are otherwise similar to phase 1: |
298 |
< |
* Retrying upon CAS misses and checking for staleness. In |
299 |
< |
* particular, if a self-link is encountered, then we can |
300 |
< |
* safely jump to a node on the list by continuing the |
301 |
< |
* traversal at current head. |
291 |
> |
* Starting at current tail pointer, find the actual last node |
292 |
> |
* and try to append a new node (or if head was null, establish |
293 |
> |
* the first node). Nodes can be appended only if their |
294 |
> |
* predecessors are either already matched or are of the same |
295 |
> |
* mode. If we detect otherwise, then a new node with opposite |
296 |
> |
* mode must have been appended during traversal, so we must |
297 |
> |
* restart at phase 1. The traversal and update steps are |
298 |
> |
* otherwise similar to phase 1: Retrying upon CAS misses and |
299 |
> |
* checking for staleness. In particular, if a self-link is |
300 |
> |
* encountered, then we can safely jump to a node on the list |
301 |
> |
* by continuing the traversal at current head. |
302 |
|
* |
303 |
< |
* On successful append, if the call was ASYNC, return |
303 |
> |
* On successful append, if the call was ASYNC, return. |
304 |
|
* |
305 |
|
* 3. Await match or cancellation (method awaitMatch) |
306 |
|
* |
307 |
|
* Wait for another thread to match node; instead cancelling if |
308 |
< |
* current thread was interrupted or the wait timed out. On |
308 |
> |
* the current thread was interrupted or the wait timed out. On |
309 |
|
* multiprocessors, we use front-of-queue spinning: If a node |
310 |
|
* appears to be the first unmatched node in the queue, it |
311 |
|
* spins a bit before blocking. In either case, before blocking |
320 |
|
* to decide to occasionally perform a Thread.yield. While |
321 |
|
* yield has underdefined specs, we assume that might it help, |
322 |
|
* and will not hurt in limiting impact of spinning on busy |
323 |
< |
* systems. We also use much smaller (1/4) spins for nodes |
324 |
< |
* that are not known to be front but whose predecessors have |
325 |
< |
* not blocked -- these "chained" spins avoid artifacts of |
323 |
> |
* systems. We also use smaller (1/2) spins for nodes that are |
324 |
> |
* not known to be front but whose predecessors have not |
325 |
> |
* blocked -- these "chained" spins avoid artifacts of |
326 |
|
* front-of-queue rules which otherwise lead to alternating |
327 |
|
* nodes spinning vs blocking. Further, front threads that |
328 |
|
* represent phase changes (from data to request node or vice |
329 |
|
* versa) compared to their predecessors receive additional |
330 |
< |
* spins, reflecting the longer code path lengths necessary to |
331 |
< |
* release them under contention. |
330 |
> |
* chained spins, reflecting longer paths typically required to |
331 |
> |
* unblock threads during phase changes. |
332 |
|
*/ |
333 |
|
|
334 |
|
/** True if on multiprocessor */ |
336 |
|
Runtime.getRuntime().availableProcessors() > 1; |
337 |
|
|
338 |
|
/** |
339 |
< |
* The number of times to spin (with on average one randomly |
340 |
< |
* interspersed call to Thread.yield) on multiprocessor before |
341 |
< |
* blocking when a node is apparently the first waiter in the |
342 |
< |
* queue. See above for explanation. Must be a power of two. The |
343 |
< |
* value is empirically derived -- it works pretty well across a |
344 |
< |
* variety of processors, numbers of CPUs, and OSes. |
339 |
> |
* The number of times to spin (with randomly interspersed calls |
340 |
> |
* to Thread.yield) on multiprocessor before blocking when a node |
341 |
> |
* is apparently the first waiter in the queue. See above for |
342 |
> |
* explanation. Must be a power of two. The value is empirically |
343 |
> |
* derived -- it works pretty well across a variety of processors, |
344 |
> |
* numbers of CPUs, and OSes. |
345 |
|
*/ |
346 |
|
private static final int FRONT_SPINS = 1 << 7; |
347 |
|
|
348 |
|
/** |
349 |
|
* The number of times to spin before blocking when a node is |
350 |
< |
* preceded by another node that is apparently spinning. |
350 |
> |
* preceded by another node that is apparently spinning. Also |
351 |
> |
* serves as an increment to FRONT_SPINS on phase changes, and as |
352 |
> |
* base average frequency for yielding during spins. Must be a |
353 |
> |
* power of two. |
354 |
|
*/ |
355 |
< |
private static final int CHAINED_SPINS = FRONT_SPINS >>> 2; |
355 |
> |
private static final int CHAINED_SPINS = FRONT_SPINS >>> 1; |
356 |
|
|
357 |
|
/** |
358 |
< |
* Queue nodes. Uses Object, not E for items to allow forgetting |
358 |
> |
* Queue nodes. Uses Object, not E, for items to allow forgetting |
359 |
|
* them after use. Relies heavily on Unsafe mechanics to minimize |
360 |
< |
* unecessary ordering constraints: Writes that intrinsically |
360 |
> |
* unnecessary ordering constraints: Writes that intrinsically |
361 |
|
* precede or follow CASes use simple relaxed forms. Other |
362 |
|
* cleanups use releasing/lazy writes. |
363 |
|
*/ |
364 |
< |
static final class Node { |
364 |
> |
static final class Node<E> { |
365 |
|
final boolean isData; // false if this is a request node |
366 |
< |
volatile Object item; // initially nonnull if isData; CASed to match |
367 |
< |
volatile Node next; |
366 |
> |
volatile Object item; // initially non-null if isData; CASed to match |
367 |
> |
volatile Node<E> next; |
368 |
|
volatile Thread waiter; // null until waiting |
369 |
|
|
370 |
|
// CAS methods for fields |
371 |
< |
final boolean casNext(Node cmp, Node val) { |
371 |
> |
final boolean casNext(Node<E> cmp, Node<E> val) { |
372 |
|
return UNSAFE.compareAndSwapObject(this, nextOffset, cmp, val); |
373 |
|
} |
374 |
|
|
375 |
|
final boolean casItem(Object cmp, Object val) { |
376 |
+ |
assert cmp == null || cmp.getClass() != Node.class; |
377 |
|
return UNSAFE.compareAndSwapObject(this, itemOffset, cmp, val); |
378 |
|
} |
379 |
|
|
380 |
|
/** |
381 |
< |
* Create a new node. Uses relaxed write because item can only |
382 |
< |
* be seen if followed by CAS |
381 |
> |
* Creates a new node. Uses relaxed write because item can only |
382 |
> |
* be seen if followed by CAS. |
383 |
|
*/ |
384 |
< |
Node(Object item, boolean isData) { |
384 |
> |
Node(E item, boolean isData) { |
385 |
|
UNSAFE.putObject(this, itemOffset, item); // relaxed write |
386 |
|
this.isData = isData; |
387 |
|
} |
410 |
|
*/ |
411 |
|
final boolean isMatched() { |
412 |
|
Object x = item; |
413 |
< |
return x == this || (x != null) != isData; |
413 |
> |
return (x == this) || ((x == null) == isData); |
414 |
|
} |
415 |
|
|
416 |
|
/** |
425 |
|
} |
426 |
|
|
427 |
|
/** |
428 |
< |
* Tries to artifically match a data node -- used by remove. |
428 |
> |
* Tries to artificially match a data node -- used by remove. |
429 |
|
*/ |
430 |
|
final boolean tryMatchData() { |
431 |
|
Object x = item; |
449 |
|
} |
450 |
|
|
451 |
|
/** head of the queue; null until first enqueue */ |
452 |
< |
private transient volatile Node head; |
452 |
> |
transient volatile Node<E> head; |
453 |
|
|
454 |
|
/** predecessor of dangling unspliceable node */ |
455 |
< |
private transient volatile Node cleanMe; // decl here to reduce contention |
455 |
> |
private transient volatile Node<E> cleanMe; // decl here reduces contention |
456 |
|
|
457 |
|
/** tail of the queue; null until first append */ |
458 |
< |
private transient volatile Node tail; |
458 |
> |
private transient volatile Node<E> tail; |
459 |
|
|
460 |
|
// CAS methods for fields |
461 |
< |
private boolean casTail(Node cmp, Node val) { |
461 |
> |
private boolean casTail(Node<E> cmp, Node<E> val) { |
462 |
|
return UNSAFE.compareAndSwapObject(this, tailOffset, cmp, val); |
463 |
|
} |
464 |
|
|
465 |
< |
private boolean casHead(Node cmp, Node val) { |
465 |
> |
private boolean casHead(Node<E> cmp, Node<E> val) { |
466 |
|
return UNSAFE.compareAndSwapObject(this, headOffset, cmp, val); |
467 |
|
} |
468 |
|
|
469 |
< |
private boolean casCleanMe(Node cmp, Node val) { |
469 |
> |
private boolean casCleanMe(Node<E> cmp, Node<E> val) { |
470 |
|
return UNSAFE.compareAndSwapObject(this, cleanMeOffset, cmp, val); |
471 |
|
} |
472 |
|
|
479 |
|
private static final int SYNC = 2; // for transfer, take |
480 |
|
private static final int TIMEOUT = 3; // for timed poll, tryTransfer |
481 |
|
|
482 |
+ |
@SuppressWarnings("unchecked") |
483 |
+ |
static <E> E cast(Object item) { |
484 |
+ |
assert item == null || item.getClass() != Node.class; |
485 |
+ |
return (E) item; |
486 |
+ |
} |
487 |
+ |
|
488 |
|
/** |
489 |
|
* Implements all queuing methods. See above for explanation. |
490 |
|
* |
491 |
|
* @param e the item or null for take |
492 |
< |
* @param haveData true if this is a put else a take |
492 |
> |
* @param haveData true if this is a put, else a take |
493 |
|
* @param how NOW, ASYNC, SYNC, or TIMEOUT |
494 |
|
* @param nanos timeout in nanosecs, used only if mode is TIMEOUT |
495 |
< |
* @return an item if matched, else e; |
495 |
> |
* @return an item if matched, else e |
496 |
|
* @throws NullPointerException if haveData mode but e is null |
497 |
|
*/ |
498 |
< |
private Object xfer(Object e, boolean haveData, int how, long nanos) { |
498 |
> |
private E xfer(E e, boolean haveData, int how, long nanos) { |
499 |
|
if (haveData && (e == null)) |
500 |
|
throw new NullPointerException(); |
501 |
< |
Node s = null; // the node to append, if needed |
501 |
> |
Node<E> s = null; // the node to append, if needed |
502 |
|
|
503 |
|
retry: for (;;) { // restart on append race |
504 |
|
|
505 |
< |
for (Node h = head, p = h; p != null;) { // find & match first node |
505 |
> |
for (Node<E> h = head, p = h; p != null;) { |
506 |
> |
// find & match first node |
507 |
|
boolean isData = p.isData; |
508 |
|
Object item = p.item; |
509 |
|
if (item != p && (item != null) == isData) { // unmatched |
510 |
|
if (isData == haveData) // can't match |
511 |
|
break; |
512 |
|
if (p.casItem(item, e)) { // match |
513 |
< |
Thread w = p.waiter; |
514 |
< |
while (p != h) { // update head |
515 |
< |
Node n = p.next; // by 2 unless singleton |
516 |
< |
if (n != null) |
517 |
< |
p = n; |
477 |
< |
if (head == h && casHead(h, p)) { |
513 |
> |
for (Node<E> q = p; q != h;) { |
514 |
> |
Node<E> n = q.next; // update head by 2 |
515 |
> |
if (n != null) // unless singleton |
516 |
> |
q = n; |
517 |
> |
if (head == h && casHead(h, q)) { |
518 |
|
h.forgetNext(); |
519 |
|
break; |
520 |
|
} // advance and retry |
521 |
|
if ((h = head) == null || |
522 |
< |
(p = h.next) == null || !p.isMatched()) |
522 |
> |
(q = h.next) == null || !q.isMatched()) |
523 |
|
break; // unless slack < 2 |
524 |
|
} |
525 |
< |
LockSupport.unpark(w); |
526 |
< |
return item; |
525 |
> |
LockSupport.unpark(p.waiter); |
526 |
> |
return this.<E>cast(item); |
527 |
|
} |
528 |
|
} |
529 |
< |
Node n = p.next; |
530 |
< |
p = p != n ? n : (h = head); // Use head if p offlist |
529 |
> |
Node<E> n = p.next; |
530 |
> |
p = (p != n) ? n : (h = head); // Use head if p offlist |
531 |
|
} |
532 |
|
|
533 |
|
if (how >= ASYNC) { // No matches available |
534 |
|
if (s == null) |
535 |
< |
s = new Node(e, haveData); |
536 |
< |
Node pred = tryAppend(s, haveData); |
535 |
> |
s = new Node<E>(e, haveData); |
536 |
> |
Node<E> pred = tryAppend(s, haveData); |
537 |
|
if (pred == null) |
538 |
|
continue retry; // lost race vs opposite mode |
539 |
|
if (how >= SYNC) |
540 |
< |
return awaitMatch(pred, s, e, how, nanos); |
540 |
> |
return awaitMatch(s, pred, e, how, nanos); |
541 |
|
} |
542 |
|
return e; // not waiting |
543 |
|
} |
544 |
|
} |
545 |
|
|
546 |
|
/** |
547 |
< |
* Tries to append node s as tail |
548 |
< |
* @param haveData true if appending in data mode |
547 |
> |
* Tries to append node s as tail. |
548 |
> |
* |
549 |
|
* @param s the node to append |
550 |
+ |
* @param haveData true if appending in data mode |
551 |
|
* @return null on failure due to losing race with append in |
552 |
|
* different mode, else s's predecessor, or s itself if no |
553 |
|
* predecessor |
554 |
|
*/ |
555 |
< |
private Node tryAppend(Node s, boolean haveData) { |
556 |
< |
for (Node t = tail, p = t;;) { // move p to actual tail and append |
557 |
< |
Node n, u; // temps for reads of next & tail |
555 |
> |
private Node<E> tryAppend(Node<E> s, boolean haveData) { |
556 |
> |
for (Node<E> t = tail, p = t;;) { // move p to last node and append |
557 |
> |
Node<E> n, u; // temps for reads of next & tail |
558 |
|
if (p == null && (p = head) == null) { |
559 |
|
if (casHead(null, s)) |
560 |
|
return s; // initialize |
561 |
|
} |
562 |
|
else if (p.cannotPrecede(haveData)) |
563 |
|
return null; // lost race vs opposite mode |
564 |
< |
else if ((n = p.next) != null) // Not tail; keep traversing |
564 |
> |
else if ((n = p.next) != null) // not last; keep traversing |
565 |
|
p = p != t && t != (u = tail) ? (t = u) : // stale tail |
566 |
< |
p != n ? n : null; // restart if off list |
566 |
> |
(p != n) ? n : null; // restart if off list |
567 |
|
else if (!p.casNext(null, s)) |
568 |
|
p = p.next; // re-read on CAS failure |
569 |
|
else { |
570 |
< |
if (p != t) { // Update if slack now >= 2 |
570 |
> |
if (p != t) { // update if slack now >= 2 |
571 |
|
while ((tail != t || !casTail(t, s)) && |
572 |
|
(t = tail) != null && |
573 |
|
(s = t.next) != null && // advance and retry |
581 |
|
/** |
582 |
|
* Spins/yields/blocks until node s is matched or caller gives up. |
583 |
|
* |
543 |
– |
* @param pred the predecessor of s or s or null if none |
584 |
|
* @param s the waiting node |
585 |
+ |
* @param pred the predecessor of s, or s itself if it has no |
586 |
+ |
* predecessor, or null if unknown (the null case does not occur |
587 |
+ |
* in any current calls but may in possible future extensions) |
588 |
|
* @param e the comparison value for checking match |
589 |
|
* @param how either SYNC or TIMEOUT |
590 |
|
* @param nanos timeout value |
591 |
|
* @return matched item, or e if unmatched on interrupt or timeout |
592 |
|
*/ |
593 |
< |
private Object awaitMatch(Node pred, Node s, Object e, |
551 |
< |
int how, long nanos) { |
593 |
> |
private E awaitMatch(Node<E> s, Node<E> pred, E e, int how, long nanos) { |
594 |
|
long lastTime = (how == TIMEOUT) ? System.nanoTime() : 0L; |
595 |
|
Thread w = Thread.currentThread(); |
596 |
|
int spins = -1; // initialized after first item and cancel checks |
599 |
|
for (;;) { |
600 |
|
Object item = s.item; |
601 |
|
if (item != e) { // matched |
602 |
+ |
assert item != s; |
603 |
|
s.forgetContents(); // avoid garbage |
604 |
< |
return item; |
604 |
> |
return this.<E>cast(item); |
605 |
|
} |
606 |
|
if ((w.isInterrupted() || (how == TIMEOUT && nanos <= 0)) && |
607 |
< |
s.casItem(e, s)) { // cancel |
607 |
> |
s.casItem(e, s)) { // cancel |
608 |
|
unsplice(pred, s); |
609 |
|
return e; |
610 |
|
} |
613 |
|
if ((spins = spinsFor(pred, s.isData)) > 0) |
614 |
|
randomYields = ThreadLocalRandom.current(); |
615 |
|
} |
616 |
< |
else if (spins > 0) { // spin, occasionally yield |
617 |
< |
if (randomYields.nextInt(FRONT_SPINS) == 0) |
618 |
< |
Thread.yield(); |
619 |
< |
--spins; |
616 |
> |
else if (spins > 0) { // spin |
617 |
> |
if (--spins == 0) |
618 |
> |
shortenHeadPath(); // reduce slack before blocking |
619 |
> |
else if (randomYields.nextInt(CHAINED_SPINS) == 0) |
620 |
> |
Thread.yield(); // occasionally yield |
621 |
|
} |
622 |
|
else if (s.waiter == null) { |
623 |
< |
shortenHeadPath(); // reduce slack before blocking |
580 |
< |
s.waiter = w; // request unpark |
623 |
> |
s.waiter = w; // request unpark then recheck |
624 |
|
} |
625 |
|
else if (how == TIMEOUT) { |
626 |
|
long now = System.nanoTime(); |
630 |
|
} |
631 |
|
else { |
632 |
|
LockSupport.park(this); |
633 |
+ |
s.waiter = null; |
634 |
|
spins = -1; // spin if front upon wakeup |
635 |
|
} |
636 |
|
} |
637 |
|
} |
638 |
|
|
639 |
|
/** |
640 |
< |
* Return spin/yield value for a node with given predecessor and |
640 |
> |
* Returns spin/yield value for a node with given predecessor and |
641 |
|
* data mode. See above for explanation. |
642 |
|
*/ |
643 |
< |
private static int spinsFor(Node pred, boolean haveData) { |
643 |
> |
private static int spinsFor(Node<?> pred, boolean haveData) { |
644 |
|
if (MP && pred != null) { |
645 |
< |
boolean predData = pred.isData; |
646 |
< |
if (predData != haveData) // front and phase change |
647 |
< |
return FRONT_SPINS + (FRONT_SPINS >>> 1); |
604 |
< |
if (predData != (pred.item != null)) // probably at front |
645 |
> |
if (pred.isData != haveData) // phase change |
646 |
> |
return FRONT_SPINS + CHAINED_SPINS; |
647 |
> |
if (pred.isMatched()) // probably at front |
648 |
|
return FRONT_SPINS; |
649 |
|
if (pred.waiter == null) // pred apparently spinning |
650 |
|
return CHAINED_SPINS; |
657 |
|
* or trailing node; failing on contention. |
658 |
|
*/ |
659 |
|
private void shortenHeadPath() { |
660 |
< |
Node h, hn, p, q; |
660 |
> |
Node<E> h, hn, p, q; |
661 |
|
if ((p = h = head) != null && h.isMatched() && |
662 |
|
(q = hn = h.next) != null) { |
663 |
< |
Node n; |
663 |
> |
Node<E> n; |
664 |
|
while ((n = q.next) != q) { |
665 |
|
if (n == null || !q.isMatched()) { |
666 |
|
if (hn != q && h.next == hn) |
676 |
|
/* -------------- Traversal methods -------------- */ |
677 |
|
|
678 |
|
/** |
679 |
< |
* Return the first unmatched node of the given mode, or null if |
679 |
> |
* Returns the first unmatched node of the given mode, or null if |
680 |
|
* none. Used by methods isEmpty, hasWaitingConsumer. |
681 |
|
*/ |
682 |
< |
private Node firstOfMode(boolean data) { |
683 |
< |
for (Node p = head; p != null; ) { |
682 |
> |
private Node<E> firstOfMode(boolean data) { |
683 |
> |
for (Node<E> p = head; p != null; ) { |
684 |
|
if (!p.isMatched()) |
685 |
< |
return p.isData == data? p : null; |
686 |
< |
Node n = p.next; |
687 |
< |
p = n != p ? n : head; |
685 |
> |
return (p.isData == data) ? p : null; |
686 |
> |
Node<E> n = p.next; |
687 |
> |
p = (n != p) ? n : head; |
688 |
|
} |
689 |
|
return null; |
690 |
|
} |
691 |
|
|
692 |
|
/** |
693 |
|
* Returns the item in the first unmatched node with isData; or |
694 |
< |
* null if none. Used by peek. |
694 |
> |
* null if none. Used by peek. |
695 |
|
*/ |
696 |
< |
private Object firstDataItem() { |
697 |
< |
for (Node p = head; p != null; ) { |
696 |
> |
private E firstDataItem() { |
697 |
> |
for (Node<E> p = head; p != null; ) { |
698 |
|
boolean isData = p.isData; |
699 |
|
Object item = p.item; |
700 |
|
if (item != p && (item != null) == isData) |
701 |
< |
return isData ? item : null; |
702 |
< |
Node n = p.next; |
703 |
< |
p = n != p ? n : head; |
701 |
> |
return isData ? this.<E>cast(item) : null; |
702 |
> |
Node<E> n = p.next; |
703 |
> |
p = (n != p) ? n : head; |
704 |
|
} |
705 |
|
return null; |
706 |
|
} |
707 |
|
|
708 |
|
/** |
709 |
< |
* Traverse and count nodes of the given mode. |
710 |
< |
* Used by methds size and getWaitingConsumerCount. |
709 |
> |
* Traverses and counts unmatched nodes of the given mode. |
710 |
> |
* Used by methods size and getWaitingConsumerCount. |
711 |
|
*/ |
712 |
|
private int countOfMode(boolean data) { |
713 |
|
int count = 0; |
714 |
< |
for (Node p = head; p != null; ) { |
714 |
> |
for (Node<E> p = head; p != null; ) { |
715 |
|
if (!p.isMatched()) { |
716 |
|
if (p.isData != data) |
717 |
|
return 0; |
718 |
|
if (++count == Integer.MAX_VALUE) // saturated |
719 |
|
break; |
720 |
|
} |
721 |
< |
Node n = p.next; |
721 |
> |
Node<E> n = p.next; |
722 |
|
if (n != p) |
723 |
|
p = n; |
724 |
|
else { |
730 |
|
} |
731 |
|
|
732 |
|
final class Itr implements Iterator<E> { |
733 |
< |
private Node nextNode; // next node to return item for |
734 |
< |
private Object nextItem; // the corresponding item |
735 |
< |
private Node lastRet; // last returned node, to support remove |
733 |
> |
private Node<E> nextNode; // next node to return item for |
734 |
> |
private E nextItem; // the corresponding item |
735 |
> |
private Node<E> lastRet; // last returned node, to support remove |
736 |
|
|
737 |
|
/** |
738 |
|
* Moves to next node after prev, or first node if prev null. |
739 |
|
*/ |
740 |
< |
private void advance(Node prev) { |
740 |
> |
private void advance(Node<E> prev) { |
741 |
|
lastRet = prev; |
742 |
< |
Node p; |
742 |
> |
Node<E> p; |
743 |
|
if (prev == null || (p = prev.next) == prev) |
744 |
|
p = head; |
745 |
|
while (p != null) { |
746 |
|
Object item = p.item; |
747 |
|
if (p.isData) { |
748 |
|
if (item != null && item != p) { |
749 |
< |
nextItem = item; |
749 |
> |
nextItem = LinkedTransferQueue.this.<E>cast(item); |
750 |
|
nextNode = p; |
751 |
|
return; |
752 |
|
} |
753 |
|
} |
754 |
|
else if (item == null) |
755 |
|
break; |
756 |
< |
Node n = p.next; |
757 |
< |
p = n != p ? n : head; |
756 |
> |
Node<E> n = p.next; |
757 |
> |
p = (n != p) ? n : head; |
758 |
|
} |
759 |
|
nextNode = null; |
760 |
|
} |
768 |
|
} |
769 |
|
|
770 |
|
public final E next() { |
771 |
< |
Node p = nextNode; |
771 |
> |
Node<E> p = nextNode; |
772 |
|
if (p == null) throw new NoSuchElementException(); |
773 |
< |
Object e = nextItem; |
773 |
> |
E e = nextItem; |
774 |
|
advance(p); |
775 |
< |
return (E) e; |
775 |
> |
return e; |
776 |
|
} |
777 |
|
|
778 |
|
public final void remove() { |
779 |
< |
Node p = lastRet; |
779 |
> |
Node<E> p = lastRet; |
780 |
|
if (p == null) throw new IllegalStateException(); |
781 |
|
lastRet = null; |
782 |
|
findAndRemoveNode(p); |
792 |
|
* @param pred predecessor of node to be unspliced |
793 |
|
* @param s the node to be unspliced |
794 |
|
*/ |
795 |
< |
private void unsplice(Node pred, Node s) { |
795 |
> |
private void unsplice(Node<E> pred, Node<E> s) { |
796 |
|
s.forgetContents(); // clear unneeded fields |
797 |
|
/* |
798 |
|
* At any given time, exactly one node on list cannot be |
799 |
< |
* deleted -- the last inserted node. To accommodate this, if |
800 |
< |
* we cannot delete s, we save its predecessor as "cleanMe", |
799 |
> |
* unlinked -- the last inserted node. To accommodate this, if |
800 |
> |
* we cannot unlink s, we save its predecessor as "cleanMe", |
801 |
|
* processing the previously saved version first. Because only |
802 |
|
* one node in the list can have a null next, at least one of |
803 |
|
* node s or the node previously saved can always be |
805 |
|
*/ |
806 |
|
if (pred != null && pred != s) { |
807 |
|
while (pred.next == s) { |
808 |
< |
Node oldpred = cleanMe == null? null : reclean(); |
809 |
< |
Node n = s.next; |
808 |
> |
Node<E> oldpred = (cleanMe == null) ? null : reclean(); |
809 |
> |
Node<E> n = s.next; |
810 |
|
if (n != null) { |
811 |
|
if (n != s) |
812 |
|
pred.casNext(s, n); |
825 |
|
* |
826 |
|
* @return current cleanMe node (or null) |
827 |
|
*/ |
828 |
< |
private Node reclean() { |
828 |
> |
private Node<E> reclean() { |
829 |
|
/* |
830 |
|
* cleanMe is, or at one time was, predecessor of a cancelled |
831 |
|
* node s that was the tail so could not be unspliced. If it |
836 |
|
* we can (must) clear cleanMe without unsplicing. This can |
837 |
|
* loop only due to contention. |
838 |
|
*/ |
839 |
< |
Node pred; |
839 |
> |
Node<E> pred; |
840 |
|
while ((pred = cleanMe) != null) { |
841 |
< |
Node s = pred.next; |
842 |
< |
Node n; |
841 |
> |
Node<E> s = pred.next; |
842 |
> |
Node<E> n; |
843 |
|
if (s == null || s == pred || !s.isMatched()) |
844 |
|
casCleanMe(pred, null); // already gone |
845 |
|
else if ((n = s.next) != null) { |
857 |
|
* Main implementation of Iterator.remove(). Find |
858 |
|
* and unsplice the given node. |
859 |
|
*/ |
860 |
< |
final void findAndRemoveNode(Node s) { |
860 |
> |
final void findAndRemoveNode(Node<E> s) { |
861 |
|
if (s.tryMatchData()) { |
862 |
< |
Node pred = null; |
863 |
< |
Node p = head; |
862 |
> |
Node<E> pred = null; |
863 |
> |
Node<E> p = head; |
864 |
|
while (p != null) { |
865 |
|
if (p == s) { |
866 |
|
unsplice(pred, p); |
882 |
|
*/ |
883 |
|
private boolean findAndRemove(Object e) { |
884 |
|
if (e != null) { |
885 |
< |
Node pred = null; |
886 |
< |
Node p = head; |
885 |
> |
Node<E> pred = null; |
886 |
> |
Node<E> p = head; |
887 |
|
while (p != null) { |
888 |
|
Object item = p.item; |
889 |
|
if (p.isData) { |
1032 |
|
} |
1033 |
|
|
1034 |
|
public E take() throws InterruptedException { |
1035 |
< |
Object e = xfer(null, false, SYNC, 0); |
1035 |
> |
E e = xfer(null, false, SYNC, 0); |
1036 |
|
if (e != null) |
1037 |
< |
return (E)e; |
1037 |
> |
return e; |
1038 |
|
Thread.interrupted(); |
1039 |
|
throw new InterruptedException(); |
1040 |
|
} |
1041 |
|
|
1042 |
|
public E poll(long timeout, TimeUnit unit) throws InterruptedException { |
1043 |
< |
Object e = xfer(null, false, TIMEOUT, unit.toNanos(timeout)); |
1043 |
> |
E e = xfer(null, false, TIMEOUT, unit.toNanos(timeout)); |
1044 |
|
if (e != null || !Thread.interrupted()) |
1045 |
< |
return (E)e; |
1045 |
> |
return e; |
1046 |
|
throw new InterruptedException(); |
1047 |
|
} |
1048 |
|
|
1049 |
|
public E poll() { |
1050 |
< |
return (E)xfer(null, false, NOW, 0); |
1050 |
> |
return xfer(null, false, NOW, 0); |
1051 |
|
} |
1052 |
|
|
1053 |
|
/** |
1104 |
|
} |
1105 |
|
|
1106 |
|
public E peek() { |
1107 |
< |
return (E) firstDataItem(); |
1107 |
> |
return firstDataItem(); |
1108 |
|
} |
1109 |
|
|
1110 |
|
/** |
1167 |
|
} |
1168 |
|
|
1169 |
|
/** |
1170 |
< |
* Save the state to a stream (that is, serialize it). |
1170 |
> |
* Saves the state to a stream (that is, serializes it). |
1171 |
|
* |
1172 |
|
* @serialData All of the elements (each an {@code E}) in |
1173 |
|
* the proper order, followed by a null |
1183 |
|
} |
1184 |
|
|
1185 |
|
/** |
1186 |
< |
* Reconstitute the Queue instance from a stream (that is, |
1187 |
< |
* deserialize it). |
1186 |
> |
* Reconstitutes the Queue instance from a stream (that is, |
1187 |
> |
* deserializes it). |
1188 |
|
* |
1189 |
|
* @param s the stream |
1190 |
|
*/ |
1200 |
|
} |
1201 |
|
} |
1202 |
|
|
1160 |
– |
|
1203 |
|
// Unsafe mechanics |
1204 |
|
|
1205 |
|
private static final sun.misc.Unsafe UNSAFE = getUnsafe(); |
1222 |
|
} |
1223 |
|
} |
1224 |
|
|
1225 |
< |
private static sun.misc.Unsafe getUnsafe() { |
1225 |
> |
/** |
1226 |
> |
* Returns a sun.misc.Unsafe. Suitable for use in a 3rd party package. |
1227 |
> |
* Replace with a simple call to Unsafe.getUnsafe when integrating |
1228 |
> |
* into a jdk. |
1229 |
> |
* |
1230 |
> |
* @return a sun.misc.Unsafe |
1231 |
> |
*/ |
1232 |
> |
static sun.misc.Unsafe getUnsafe() { |
1233 |
|
try { |
1234 |
|
return sun.misc.Unsafe.getUnsafe(); |
1235 |
|
} catch (SecurityException se) { |