usync 0.2.1

fast, drop-in, synchronization primitives
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
use super::{
    shared::{fence_acquire, invalid_mut, StrictProvenance, Waiter},
    MutexGuard, RawRwLock,
};
use lock_api::RawMutex as _RawMutex;
use std::{
    fmt,
    pin::Pin,
    ptr::NonNull,
    sync::atomic::{AtomicPtr, Ordering},
    time::{Duration, Instant},
};

const EMPTY: usize = 0;
const SIGNAL: usize = 1;
const SIGNAL_MASK: usize = 0b111;
const QUEUE_LOCKED: usize = SIGNAL_MASK + 1;

/// A Condition Variable
///
/// Condition variables represent the ability to block a thread such that it
/// consumes no CPU time while waiting for an event to occur. Condition
/// variables are typically associated with a boolean predicate (a condition)
/// and a mutex. The predicate is always verified inside of the mutex before
/// determining that thread must block.
///
/// Note that each condvar can be used with multiple mutexes at a time.
/// This is different compared to `parking_lot`'s condvar, which doesn't allow
/// multiple threads waiting on the same condvar with different mutexes.
///
/// # Differences from the standard library `Condvar`
///
/// - Spurious wake ups are avoided. This means a wait will try to not return early
///   unless woken up from a call to `notify_one` or `notify_all`. This is not fool-proof
///   as events such as timeouts or stacked `notify_one` calls could cause a spurious wake up to occur.
/// - `Condvar::notify_all` will try to only wake up a single thread with the rest being
///   requeued to wait for the `Mutex` to be unlocked by the thread that was
///   woken up.
/// - Only requires 1 word of space, whereas the standard library boxes the
///   `Condvar` due to platform limitations.
/// - Can be statically constructed (requires the `const_fn` nightly feature).
/// - Does not require any drop glue when dropped.
/// - Inline fast path for the uncontended case.
///
/// # Examples
///
/// ```
/// use usync::{Mutex, Condvar};
/// use std::sync::Arc;
/// use std::thread;
///
/// let pair = Arc::new((Mutex::new(false), Condvar::new()));
/// let pair2 = pair.clone();
///
/// // Inside of our lock, spawn a new thread, and then wait for it to start
/// thread::spawn(move|| {
///     let &(ref lock, ref cvar) = &*pair2;
///     let mut started = lock.lock();
///     *started = true;
///     cvar.notify_one();
/// });
///
/// // wait for the thread to start up
/// let &(ref lock, ref cvar) = &*pair;
/// let mut started = lock.lock();
/// if !*started {
///     cvar.wait(&mut started);
/// }
/// // Note that we used an if instead of a while loop above. This is only
/// // possible because usync's Condvar will only spuriously wake up when timeouts
/// // are involved. This means that wait() will only return after notify_one or
/// // notify_all is called in this instance.
/// ```
#[derive(Default)]
pub struct Condvar {
    /// This atomic integer holds the current state of the Condvar instance.
    /// The four least significant bits are used to notify_one calls and the state of the Condvar.
    ///
    /// # State table:
    ///
    /// SIGNAL_MASK | QUEUE_LOCKED | Remaining | Description
    ///       0     |       0       |     0     | The condvar is empty and there's nothing waiting on it.
    /// ------------+---------------+-----------+-----------------------------------------------------------------
    ///       0     |       0       |  *Waiter  | The Remaining bits point to the head Waiter node of the
    ///             |               |           | waiting thread queue.
    /// ------------+---------------+-----------+-----------------------------------------------------------------
    ///       0     |       1       |  *Waiter  | The Remaining bits point to the head Waiter node of the
    ///             |               |           | waiting thread queue. There is also a thread which is
    ///             |               |           | updating the waiting-thread queue and possibly waking from it.
    /// ------------+---------------+-----------+-----------------------------------------------------------------
    ///      n>0    |       1       |  *Waiter  | The Remaining bits point to the head Waiter node of the
    ///             |               |           | waiting thread queue. There is also a thread which is
    ///             |               |           | updating the waiting-thread queue and waking from it.
    ///             |               |           | `n` is the amount of `notify_one` calls that happened since the
    ///             |               |           | QUEUE_LOCKED bit was taken for waiting-thread queue updating/waking.
    //              |               |           | If `n == SIGNAL_MASK` then all threads in the queue are woken up.
    /// ------------+---------------+-----------+-----------------------------------------------------------------
    state: AtomicPtr<Waiter>,
}

impl fmt::Debug for Condvar {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        f.pad("Condvar { .. }")
    }
}

unsafe impl Send for Condvar {}
unsafe impl Sync for Condvar {}

impl Condvar {
    /// Creates a new condition variable which is ready to be waited on and
    /// notified.
    pub const fn new() -> Self {
        Self {
            state: AtomicPtr::new(invalid_mut(EMPTY)),
        }
    }

    /// Blocks the current thread until this condition variable receives a
    /// notification.
    ///
    /// This function will atomically unlock the mutex specified (represented by
    /// `mutex_guard`) and block the current thread. This means that any calls
    /// to `notify_*()` which happen logically after the mutex is unlocked are
    /// candidates to wake this thread up. When this function call returns, the
    /// lock specified will have been re-acquired.
    pub fn wait<T: ?Sized>(&self, mutex_guard: &mut MutexGuard<'_, T>) {
        let result = self.wait_with(mutex_guard, None);
        assert!(!result.timed_out());
    }

    /// Waits on this condition variable for a notification, timing out after
    /// the specified time instant.
    ///
    /// The semantics of this function are equivalent to `wait()` except that
    /// the thread will be blocked roughly until `timeout` is reached. This
    /// method should not be used for precise timing due to anomalies such as
    /// preemption or platform differences that may not cause the maximum
    /// amount of time waited to be precisely `timeout`.
    ///
    /// Note that the best effort is made to ensure that the time waited is
    /// measured with a monotonic clock, and not affected by the changes made to
    /// the system time.
    ///
    /// The returned `WaitTimeoutResult` value indicates if the timeout is
    /// known to have elapsed.
    ///
    /// Like `wait`, the lock specified will be re-acquired when this function
    /// returns, regardless of whether the timeout elapsed or not.
    pub fn wait_until<T: ?Sized>(
        &self,
        mutex_guard: &mut MutexGuard<'_, T>,
        timeout: Instant,
    ) -> WaitTimeoutResult {
        match timeout.checked_duration_since(Instant::now()) {
            Some(until_deadline) => self.wait_for(mutex_guard, until_deadline),
            None => WaitTimeoutResult(true),
        }
    }

    /// Waits on this condition variable for a notification, timing out after a
    /// specified duration.
    ///
    /// The semantics of this function are equivalent to `wait()` except that
    /// the thread will be blocked for roughly no longer than `timeout`. This
    /// method should not be used for precise timing due to anomalies such as
    /// preemption or platform differences that may not cause the maximum
    /// amount of time waited to be precisely `timeout`.
    ///
    /// Note that the best effort is made to ensure that the time waited is
    /// measured with a monotonic clock, and not affected by the changes made to
    /// the system time.
    ///
    /// The returned `WaitTimeoutResult` value indicates if the timeout is
    /// known to have elapsed.
    ///
    /// Like `wait`, the lock specified will be re-acquired when this function
    /// returns, regardless of whether the timeout elapsed or not.
    pub fn wait_for<T: ?Sized>(
        &self,
        mutex_guard: &mut MutexGuard<'_, T>,
        timeout: Duration,
    ) -> WaitTimeoutResult {
        self.wait_with(mutex_guard, Some(timeout))
    }

    #[cold]
    fn wait_with<T: ?Sized>(
        &self,
        mutex_guard: &mut MutexGuard<'_, T>,
        timeout: Option<Duration>,
    ) -> WaitTimeoutResult {
        Waiter::with(|waiter| unsafe {
            // MutexGuard acquired the internal RawRwLock as a writer
            let is_writer = true;
            waiter.flags.set(is_writer as usize);

            // RawMutex is just a wrapper around RawRwLock.
            let raw_mutex = MutexGuard::mutex(mutex_guard).raw();
            let raw_rwlock = NonNull::from(&raw_mutex.rwlock);

            waiter.waiting_on.set(Some(raw_rwlock.cast()));
            waiter.prev.set(None);

            // Push our waiter to the wait queue and report if we acquired the QUEUE_LOCKED bit
            let mut state = self.state.load(Ordering::Relaxed);
            let signal_locked = loop {
                let head = NonNull::new(state.map_address(|addr| addr & Waiter::MASK));
                waiter.next.set(head);

                let waiter_ptr = NonNull::from(&*waiter).as_ptr();
                let mut new_state = waiter_ptr.map_address(|addr| {
                    let state_bits = state.address() & !Waiter::MASK;
                    addr | state_bits
                });

                // If we're the first waiter, set our tail to be ourselves.
                // This allows Waiter::get_and_link_queue() to found our tail.
                // If we're not the first waiter, try to acquire the QUEUE_LOCKED bit to link the queue.
                if head.is_some() {
                    waiter.tail.set(None);
                    new_state = new_state.map_address(|addr| addr | QUEUE_LOCKED);
                } else {
                    waiter.tail.set(Some(NonNull::from(&*waiter)));
                }

                // Release barrier to ensure that our waiter writes above
                // happen before the QUEUE_LOCKED bit holder operates on the queue visible from state.
                match self.state.compare_exchange_weak(
                    state,
                    new_state,
                    Ordering::Release,
                    Ordering::Relaxed,
                ) {
                    Ok(_) => break head.is_some() && (state.address() & QUEUE_LOCKED == 0),
                    Err(e) => state = e,
                }
            };

            struct DropGuard<'a>(&'a crate::RawMutex);
            impl<'a> Drop for DropGuard<'a> {
                fn drop(&mut self) {
                    self.0.lock();
                }
            }

            // Now that our waiter is registered on the state, unlock the mutex in order to block the thread.
            // Make sure to re-acquire the mutex back when returning (even in the case of a panic).
            raw_mutex.unlock();
            let _drop_guard = DropGuard(raw_mutex);

            // Make sure to link and unset the QUEUE_LOCKED
            if signal_locked {
                state = self.state.load(Ordering::Relaxed);
                self.link_queue_or_unpark(state);
            }

            // Block the thread and wait for a wake up or timeout.
            let timed_out = !waiter.parker.park(timeout);

            // On timeout, we must ensure that our waiter is no longer in the waiting-thread queue.
            // We could try to grab the QUEUE_LOCKED bit and remove ourselves, but it's not guaranteed
            // that we're still in the waiting-thread queue; We could have been requeued to the RawRWLock.
            // Instead, just wake everything up and wait for our waiter specifically to be woken up.
            if timed_out {
                self.notify_all();
                assert!(waiter.parker.park(None));
            }

            // return whether we timed out (will re-acquire the mutex when returning)
            WaitTimeoutResult(timed_out)
        })
    }

    #[cold]
    unsafe fn link_queue_or_unpark(&self, mut state: *mut Waiter) {
        loop {
            assert_ne!(state.address() & QUEUE_LOCKED, 0);

            // If `notify_*` calls occured while we we're trying to link the queue,
            // then we are now responsible for doing the wake up from the notifications.
            let signals = state.address() & SIGNAL_MASK;
            if signals > 0 {
                return self.unpark(state, 0);
            }

            // Fix the prev links in the waiter queue now that we hold the QUEUE_LOCKED bit.
            // Acquire barrier to ensure writes to waiters pushed to the queue happen before we start fixing it.
            fence_acquire(&self.state);
            let _ = Waiter::get_and_link_queue(state, |_| {});

            // Finally, try to the release the QUEUE_LOCKED bit.
            // Release barrier to ensure the writes we did above happen before the next QUEUE_LOCKED bit holder.
            match self.state.compare_exchange_weak(
                state,
                state.map_address(|addr| addr & !QUEUE_LOCKED),
                Ordering::Release,
                Ordering::Relaxed,
            ) {
                Ok(_) => return,
                Err(e) => state = e,
            }
        }
    }

    /// Wakes up one blocked thread on this condvar.
    ///
    /// Returns **a hint** as to whether a thread was woken up.
    ///
    /// If there is a blocked thread on this condition variable, then it will
    /// be woken up from its call to `wait` or `wait_timeout`. Calls to
    /// `notify_one` are not buffered in any way *to subsequent waiters*.
    ///
    /// To wake up all threads, see `notify_all()`.
    ///
    /// # Examples
    ///
    /// ```
    /// use usync::Condvar;
    ///
    /// let condvar = Condvar::new();
    ///
    /// // do something with condvar, share it with other threads
    ///
    /// if !condvar.notify_one() {
    ///     println!("Nobody was listening for this.");
    /// }
    /// ```
    #[inline]
    pub fn notify_one(&self) -> bool {
        let state = self.state.load(Ordering::Relaxed);
        if state.address() == EMPTY {
            return false;
        }

        self.notify_one_slow(state)
    }

    #[cold]
    fn notify_one_slow(&self, mut state: *mut Waiter) -> bool {
        loop {
            if state.address() == EMPTY {
                return false;
            }

            if state.address() & QUEUE_LOCKED == 0 {
                // Try to acquire the QUEUE_LOCKED bit to wake up the queued threads.
                let new_state = state.map_address(|addr| addr | QUEUE_LOCKED);
                if let Err(e) = self.state.compare_exchange_weak(
                    state,
                    new_state,
                    Ordering::Relaxed,
                    Ordering::Relaxed,
                ) {
                    state = e;
                    continue;
                }

                unsafe { self.unpark(new_state, 1) };
                return true;
            }

            // Bail if all threads are going to be woken up eventually.
            let signals = state.address() & SIGNAL_MASK;
            if signals == SIGNAL_MASK {
                return false;
            }

            // Tell the QUEUE_LOCKED bit holder to wake up the thread in our place.
            // Release barrier to ensure this notify_one() happens before
            // the wake ups done by the QUEUE_LOCKED bit holder.
            match self.state.compare_exchange_weak(
                state,
                state.map_address(|addr| addr + SIGNAL),
                Ordering::Release,
                Ordering::Relaxed,
            ) {
                Ok(_) => return true,
                Err(e) => state = e,
            }
        }
    }

    /// Wakes up all blocked threads on this condvar.
    ///
    /// Returns **a hint** as to whether threads were woken up.
    ///
    /// This method will ensure that any current waiters on the condition
    /// variable are awoken. Calls to `notify_all()` are not buffered in any
    /// way *to subsequent waiters*.
    ///
    /// To wake up only one thread, see `notify_one()`.
    #[inline]
    pub fn notify_all(&self) -> bool {
        let state = self.state.load(Ordering::Relaxed);
        if state.address() == EMPTY {
            return false;
        }

        self.notify_all_slow(state)
    }

    #[cold]
    fn notify_all_slow(&self, mut state: *mut Waiter) -> bool {
        loop {
            if state.address() == EMPTY {
                return false;
            }

            // If no thread is currently waking up the waiters, grab all of them to wake up.
            // Acquire barrier to ensure writes to pushed waiters happen before we start waking them up.
            if state.address() & QUEUE_LOCKED == 0 {
                if let Err(e) = self.state.compare_exchange_weak(
                    state,
                    state.with_address(EMPTY),
                    Ordering::Acquire,
                    Ordering::Relaxed,
                ) {
                    state = e;
                    continue;
                }

                unsafe { self.unpark_waiters(state) };
                return true;
            }

            let signals = state.address() & SIGNAL_MASK;
            if signals == SIGNAL_MASK {
                return false;
            }

            // Tell the QUEUE_LOCKED bit holder to wake up all threads in our place.
            // Release barrier to ensure this notify_one() happens before
            // the wake ups done by the QUEUE_LOCKED bit holder.
            match self.state.compare_exchange_weak(
                state,
                state.map_address(|addr| addr | SIGNAL_MASK),
                Ordering::Release,
                Ordering::Relaxed,
            ) {
                Ok(_) => return true,
                Err(e) => state = e,
            }
        }
    }

    #[cold]
    unsafe fn unpark(&self, mut state: *mut Waiter, also_wake: usize) {
        let mut scanned = 0;
        let mut unparked = None;

        loop {
            assert_ne!(state.address() & QUEUE_LOCKED, 0);

            // If enough threads have sent notifications, just wake up all waiters.
            let signals = state.address() & SIGNAL_MASK;
            if signals == SIGNAL_MASK {
                return self.unpark_all();
            }

            // Fix and get the ends of the wait queue in order to wake up waiters starting from the tail.
            // Acquire barrier ensures that writes to waiters pushed to the queue
            // happen before we start fixing/getting it.
            fence_acquire(&self.state);
            let (head, tail) = Waiter::get_and_link_queue(state, |_| {});

            // Get the head of the waiter nodes we plan on waking up.
            let mut front = unparked.unwrap_or_else(|| {
                scanned += 1;
                tail
            });

            // Bounded scan to collect enough waiter nodes to match current buffered signals.
            let max_scan = signals + also_wake;
            while scanned < max_scan {
                if let Some(prev) = front.as_ref().prev.get() {
                    front = prev;
                    scanned += 1;
                } else {
                    break;
                }
            }

            // If enough threads have been scanned, just wake up all waiters.
            unparked = Some(front);
            if scanned >= SIGNAL_MASK {
                return self.unpark_all();
            }

            // If we're only waking up a portion of the queue,
            // try to modify that queue to remove said portion only.
            if let Some(new_tail) = front.as_ref().prev.get() {
                head.as_ref().tail.set(Some(new_tail));
                new_tail.as_ref().next.set(None);

                // Use saturating sub as we may have scanned to wake up a bit more than whats signaled
                // if unpark() is being called by `notify_one()` which wakes without adding a SIGNAL.
                let new_state = state.map_address(|mut addr| {
                    addr &= !QUEUE_LOCKED;
                    addr -= signals.saturating_sub(scanned) * SIGNAL;
                    addr
                });

                // Release barrier ensures the head/tail access above happen before we release the QUEUE_LOCKED bit before wake up.
                match self.state.compare_exchange_weak(
                    state,
                    new_state,
                    Ordering::Release,
                    Ordering::Relaxed,
                ) {
                    Ok(_) => return self.unpark_requeue(front),
                    Err(e) => state = e,
                }

                // Reverse wha we did above
                new_tail.as_ref().next.set(Some(front));
                head.as_ref().tail.set(Some(tail));
                continue;
            }

            // Wake all by zeroing out all the SIGNALS, unsetting the QUEUE_LOCKED bit, and unsetting the waiter queue head pointer.
            // Release barrier ensures the head/tail access above happen before we release the QUEUE_LOCKED bit before wake up.
            match self.state.compare_exchange_weak(
                state,
                state.with_address(EMPTY),
                Ordering::Release,
                Ordering::Relaxed,
            ) {
                Ok(_) => return self.unpark_requeue(head),
                Err(e) => state = e,
            }
        }
    }

    #[cold]
    unsafe fn unpark_all(&self) {
        // Wake all by zeroing everything.
        // Acquire barrier ensures that writes done to pushed waiters happen before we start waking them.
        let state = self.state.swap(invalid_mut(EMPTY), Ordering::Acquire);
        assert_ne!(state.address() & QUEUE_LOCKED, 0);
        assert_ne!(state.address() & SIGNAL_MASK, 0);

        self.unpark_waiters(state)
    }

    #[cold]
    unsafe fn unpark_waiters(&self, state: *mut Waiter) {
        // Get the head waiter node from the queue and wake the entire queue.
        let head = NonNull::new(state.map_address(|addr| addr & Waiter::MASK));
        let head = head.expect("Condvar waking up waiters with invalid state");

        self.unpark_requeue(head)
    }

    #[cold]
    unsafe fn unpark_requeue(&self, head: NonNull<Waiter>) {
        let mut waiters = Some(head);
        while let Some(waiter) = waiters {
            waiters = waiter.as_ref().next.get();

            let waiting_on = waiter.as_ref().waiting_on.get();
            let waiting_on = waiting_on.expect("Condvar waiter not waiting on anything");

            // Storing the lock the waiter is waiting on inside the waiter
            // allows the Condvar to support waiting on multiples mutexes at once.
            let raw_rwlock = waiting_on.cast::<RawRwLock>();
            let waiter = Pin::new_unchecked(waiter.as_ref());

            // Try to requeue the waiter onto the RwLock (really, Mutex) it was waiting on.
            // Failure to do so means the lock is unlocked and we should unpark directly in
            // hopes that the waiter will immediately acquire it.
            if !raw_rwlock.as_ref().try_requeue(waiter) {
                waiter.parker.unpark();
            }
        }
    }
}

/// A type indicating whether a timed wait on a condition variable returned
/// due to a time out or not.
#[derive(Debug, PartialEq, Eq, Copy, Clone)]
pub struct WaitTimeoutResult(bool);

impl WaitTimeoutResult {
    /// Returns whether the wait was known to have timed out.
    #[inline]
    pub fn timed_out(self) -> bool {
        self.0
    }
}

#[cfg(test)]
mod tests {
    use crate::{Condvar, Mutex};
    use std::{
        sync::{mpsc::channel, Arc},
        thread,
        time::{Duration, Instant},
    };

    #[test]
    fn smoke() {
        let c = Condvar::new();
        c.notify_one();
        c.notify_all();
    }

    #[test]
    fn notify_one() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        let mut g = m.lock();
        let _t = thread::spawn(move || {
            let _g = m2.lock();
            c2.notify_one();
        });
        c.wait(&mut g);
    }

    #[test]
    fn notify_all() {
        const N: usize = 10;

        let data = Arc::new((Mutex::new(0), Condvar::new()));
        let (tx, rx) = channel();
        for _ in 0..N {
            let data = data.clone();
            let tx = tx.clone();
            thread::spawn(move || {
                let &(ref lock, ref cond) = &*data;
                let mut cnt = lock.lock();
                *cnt += 1;
                if *cnt == N {
                    tx.send(()).unwrap();
                }
                while *cnt != 0 {
                    cond.wait(&mut cnt);
                }
                tx.send(()).unwrap();
            });
        }
        drop(tx);

        let &(ref lock, ref cond) = &*data;
        rx.recv().unwrap();
        let mut cnt = lock.lock();
        *cnt = 0;
        cond.notify_all();
        drop(cnt);

        for _ in 0..N {
            rx.recv().unwrap();
        }
    }

    #[test]
    fn notify_one_return_true() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        let mut g = m.lock();
        let _t = thread::spawn(move || {
            let _g = m2.lock();
            assert!(c2.notify_one());
        });
        c.wait(&mut g);
    }

    #[test]
    fn notify_one_return_false() {
        let m = Arc::new(Mutex::new(()));
        let c = Arc::new(Condvar::new());

        let _t = thread::spawn(move || {
            let _g = m.lock();
            assert!(!c.notify_one());
        });
    }

    #[test]
    fn notify_all_return() {
        const N: usize = 10;

        let data = Arc::new((Mutex::new(0), Condvar::new()));
        let (tx, rx) = channel();
        for _ in 0..N {
            let data = data.clone();
            let tx = tx.clone();
            thread::spawn(move || {
                let &(ref lock, ref cond) = &*data;
                let mut cnt = lock.lock();
                *cnt += 1;
                if *cnt == N {
                    tx.send(()).unwrap();
                }
                while *cnt != 0 {
                    cond.wait(&mut cnt);
                }
                tx.send(()).unwrap();
            });
        }
        drop(tx);

        let &(ref lock, ref cond) = &*data;
        rx.recv().unwrap();
        let mut cnt = lock.lock();
        *cnt = 0;
        assert_eq!(cond.notify_all(), true);
        drop(cnt);

        for _ in 0..N {
            rx.recv().unwrap();
        }

        assert_eq!(cond.notify_all(), false);
    }

    #[test]
    fn wait_for() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        let mut g = m.lock();
        let no_timeout = c.wait_for(&mut g, Duration::from_millis(1));
        assert!(no_timeout.timed_out());

        let _t = thread::spawn(move || {
            let _g = m2.lock();
            c2.notify_one();
        });
        let timeout_res = c.wait_for(&mut g, Duration::from_secs(u64::max_value()));
        assert!(!timeout_res.timed_out());

        drop(g);
    }

    #[test]
    fn wait_until() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        let mut g = m.lock();
        let no_timeout = c.wait_until(&mut g, Instant::now() + Duration::from_millis(1));
        assert!(no_timeout.timed_out());
        let _t = thread::spawn(move || {
            let _g = m2.lock();
            c2.notify_one();
        });
        let timeout_res = c.wait_until(
            &mut g,
            Instant::now() + Duration::from_millis(u32::max_value() as u64),
        );
        assert!(!timeout_res.timed_out());
        drop(g);
    }

    #[test]
    fn two_mutexes() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let m3 = Arc::new(Mutex::new(()));
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        // Make sure we don't leave the child thread dangling
        struct PanicGuard<'a>(&'a Condvar);
        impl<'a> Drop for PanicGuard<'a> {
            fn drop(&mut self) {
                self.0.notify_one();
            }
        }

        let (tx, rx) = channel();
        let g = m.lock();
        let _t = thread::spawn(move || {
            let mut g = m2.lock();
            tx.send(()).unwrap();
            c2.wait(&mut g);
        });
        drop(g);
        rx.recv().unwrap();
        let _g = m.lock();
        let _guard = PanicGuard(&*c);

        let result = c.wait_for(&mut m3.lock(), Duration::from_millis(100));
        assert!(result.timed_out());
    }

    #[test]
    fn two_mutexes_disjoint() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let m3 = Arc::new(Mutex::new(()));
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();

        let mut g = m.lock();
        let _t = thread::spawn(move || {
            let _g = m2.lock();
            c2.notify_one();
        });
        c.wait(&mut g);
        drop(g);

        let _ = c.wait_for(&mut m3.lock(), Duration::from_millis(1));
    }

    #[test]
    fn test_debug_condvar() {
        let c = Condvar::new();
        assert_eq!(format!("{:?}", c), "Condvar { .. }");
    }

    #[test]
    fn test_condvar_requeue() {
        let m = Arc::new(Mutex::new(()));
        let m2 = m.clone();
        let c = Arc::new(Condvar::new());
        let c2 = c.clone();
        let t = thread::spawn(move || {
            let mut g = m2.lock();
            c2.wait(&mut g);
        });

        let mut g = m.lock();
        while !c.notify_one() {
            // Wait for the thread to get into wait()

            //MutexGuard::bump(&mut g);
            drop(g);
            thread::yield_now();
            g = m.lock();

            // Yield, so the other thread gets a chance to do something.
            // (At least Miri needs this, because it doesn't preempt threads.)
            thread::yield_now();
        }
        // The thread should have been requeued to the mutex, which we wake up now.
        drop(g);
        t.join().unwrap();
    }

    #[test]
    fn test_parking_lot_issue_129() {
        let locks = Arc::new((Mutex::new(()), Condvar::new()));

        let (tx, rx) = channel();
        for _ in 0..4 {
            let locks = locks.clone();
            let tx = tx.clone();
            thread::spawn(move || {
                let mut guard = locks.0.lock();
                locks.1.wait(&mut guard);
                locks.1.wait_for(&mut guard, Duration::from_millis(1));
                locks.1.notify_one();
                tx.send(()).unwrap();
            });
        }

        thread::sleep(Duration::from_millis(100));
        locks.1.notify_one();

        for _ in 0..4 {
            assert_eq!(rx.recv_timeout(Duration::from_millis(500)), Ok(()));
        }
    }
}

/// This module contains an integration test that is heavily inspired from WebKit's own integration
/// tests for it's own Condvar.
#[cfg(test)]
mod webkit_queue_test {
    use crate::{Condvar, Mutex, MutexGuard};
    use std::{collections::VecDeque, sync::Arc, thread, time::Duration};

    #[derive(Clone, Copy)]
    enum Timeout {
        Bounded(Duration),
        Forever,
    }

    #[derive(Clone, Copy)]
    enum NotifyStyle {
        One,
        All,
    }

    struct Queue {
        items: VecDeque<usize>,
        should_continue: bool,
    }

    impl Queue {
        fn new() -> Self {
            Self {
                items: VecDeque::new(),
                should_continue: true,
            }
        }
    }

    fn wait<T: ?Sized>(
        condition: &Condvar,
        lock: &mut MutexGuard<'_, T>,
        predicate: impl Fn(&mut MutexGuard<'_, T>) -> bool,
        timeout: &Timeout,
    ) {
        while !predicate(lock) {
            match timeout {
                Timeout::Forever => condition.wait(lock),
                Timeout::Bounded(bound) => {
                    condition.wait_for(lock, *bound);
                }
            }
        }
    }

    fn notify(style: NotifyStyle, condition: &Condvar, should_notify: bool) {
        match style {
            NotifyStyle::One => {
                condition.notify_one();
            }
            NotifyStyle::All => {
                if should_notify {
                    condition.notify_all();
                }
            }
        }
    }

    fn run_queue_test(
        num_producers: usize,
        num_consumers: usize,
        max_queue_size: usize,
        messages_per_producer: usize,
        notify_style: NotifyStyle,
        timeout: Timeout,
        delay: Duration,
    ) {
        let input_queue = Arc::new(Mutex::new(Queue::new()));
        let empty_condition = Arc::new(Condvar::new());
        let full_condition = Arc::new(Condvar::new());

        let output_vec = Arc::new(Mutex::new(vec![]));

        let consumers = (0..num_consumers)
            .map(|_| {
                consumer_thread(
                    input_queue.clone(),
                    empty_condition.clone(),
                    full_condition.clone(),
                    timeout,
                    notify_style,
                    output_vec.clone(),
                    max_queue_size,
                )
            })
            .collect::<Vec<_>>();
        let producers = (0..num_producers)
            .map(|_| {
                producer_thread(
                    messages_per_producer,
                    input_queue.clone(),
                    empty_condition.clone(),
                    full_condition.clone(),
                    timeout,
                    notify_style,
                    max_queue_size,
                )
            })
            .collect::<Vec<_>>();

        thread::sleep(delay);

        for producer in producers.into_iter() {
            producer.join().expect("Producer thread panicked");
        }

        {
            let mut input_queue = input_queue.lock();
            input_queue.should_continue = false;
        }
        empty_condition.notify_all();

        for consumer in consumers.into_iter() {
            consumer.join().expect("Consumer thread panicked");
        }

        let mut output_vec = output_vec.lock();
        assert_eq!(output_vec.len(), num_producers * messages_per_producer);
        output_vec.sort();
        for msg_idx in 0..messages_per_producer {
            for producer_idx in 0..num_producers {
                assert_eq!(msg_idx, output_vec[msg_idx * num_producers + producer_idx]);
            }
        }
    }

    fn consumer_thread(
        input_queue: Arc<Mutex<Queue>>,
        empty_condition: Arc<Condvar>,
        full_condition: Arc<Condvar>,
        timeout: Timeout,
        notify_style: NotifyStyle,
        output_queue: Arc<Mutex<Vec<usize>>>,
        max_queue_size: usize,
    ) -> thread::JoinHandle<()> {
        thread::spawn(move || loop {
            let (should_notify, result) = {
                let mut queue = input_queue.lock();
                wait(
                    &*empty_condition,
                    &mut queue,
                    |state| -> bool { !state.items.is_empty() || !state.should_continue },
                    &timeout,
                );
                if queue.items.is_empty() && !queue.should_continue {
                    return;
                }
                let should_notify = queue.items.len() == max_queue_size;
                let result = queue.items.pop_front();
                std::mem::drop(queue);
                (should_notify, result)
            };
            notify(notify_style, &*full_condition, should_notify);

            if let Some(result) = result {
                output_queue.lock().push(result);
            }
        })
    }

    fn producer_thread(
        num_messages: usize,
        queue: Arc<Mutex<Queue>>,
        empty_condition: Arc<Condvar>,
        full_condition: Arc<Condvar>,
        timeout: Timeout,
        notify_style: NotifyStyle,
        max_queue_size: usize,
    ) -> thread::JoinHandle<()> {
        thread::spawn(move || {
            for message in 0..num_messages {
                let should_notify = {
                    let mut queue = queue.lock();
                    wait(
                        &*full_condition,
                        &mut queue,
                        |state| state.items.len() < max_queue_size,
                        &timeout,
                    );
                    let should_notify = queue.items.is_empty();
                    queue.items.push_back(message);
                    std::mem::drop(queue);
                    should_notify
                };
                notify(notify_style, &*empty_condition, should_notify);
            }
        })
    }

    macro_rules! run_queue_tests {
        ( $( $name:ident(
            num_producers: $num_producers:expr,
            num_consumers: $num_consumers:expr,
            max_queue_size: $max_queue_size:expr,
            messages_per_producer: $messages_per_producer:expr,
            notification_style: $notification_style:expr,
            timeout: $timeout:expr,
            delay_seconds: $delay_seconds:expr);
        )* ) => {
            $(#[test]
            fn $name() {
                let delay = Duration::from_secs($delay_seconds);
                run_queue_test(
                    $num_producers,
                    $num_consumers,
                    $max_queue_size,
                    $messages_per_producer,
                    $notification_style,
                    $timeout,
                    delay,
                    );
            })*
        };
    }

    run_queue_tests! {
        sanity_check_queue(
            num_producers: 1,
            num_consumers: 1,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Bounded(Duration::from_secs(1)),
            delay_seconds: 0
        );
        sanity_check_queue_timeout(
            num_producers: 1,
            num_consumers: 1,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        new_test_without_timeout_5(
            num_producers: 1,
            num_consumers: 5,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        one_producer_one_consumer_one_slot(
            num_producers: 1,
            num_consumers: 1,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        one_producer_one_consumer_one_slot_timeout(
            num_producers: 1,
            num_consumers: 1,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 1
        );
        one_producer_one_consumer_hundred_slots(
            num_producers: 1,
            num_consumers: 1,
            max_queue_size: if cfg!(miri) { 10 } else { 100 },
            messages_per_producer: if cfg!(miri) { 100 } else { 1000_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_one_consumer_one_slot(
            num_producers: 10,
            num_consumers: 1,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 50 } else { 10000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_one_consumer_hundred_slots_notify_all(
            num_producers: 10,
            num_consumers: 1,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 50 } else { 10000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_one_consumer_hundred_slots_notify_one(
            num_producers: 10,
            num_consumers: 1,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 50 } else { 10000 },
            notification_style: NotifyStyle::One,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        one_producer_ten_consumers_one_slot(
            num_producers: 1,
            num_consumers: 10,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 50 } else { 10000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        one_producer_ten_consumers_hundred_slots_notify_all(
            num_producers: 1,
            num_consumers: 10,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        one_producer_ten_consumers_hundred_slots_notify_one(
            num_producers: 1,
            num_consumers: 10,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 100 } else { 100_000 },
            notification_style: NotifyStyle::One,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_ten_consumers_one_slot(
            num_producers: 10,
            num_consumers: 10,
            max_queue_size: 1,
            messages_per_producer: if cfg!(miri) { 50 } else { 50000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_ten_consumers_hundred_slots_notify_all(
            num_producers: 10,
            num_consumers: 10,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 50 } else { 50000 },
            notification_style: NotifyStyle::All,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
        ten_producers_ten_consumers_hundred_slots_notify_one(
            num_producers: 10,
            num_consumers: 10,
            max_queue_size: 100,
            messages_per_producer: if cfg!(miri) { 50 } else { 50000 },
            notification_style: NotifyStyle::One,
            timeout: Timeout::Forever,
            delay_seconds: 0
        );
    }
}