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
//! Multi-producer, single-consumer channels using [`ThingBuf`].
//!
//! A _channel_ is a synchronization and communication primitive that combines a
//! shared queue with the ability to _wait_. Channels provide a [send]
//! operation, which enqueues a message if there is capacity in the queue, or
//! waits for capacity to become available if there is none; and a [receive]
//! operation, which dequeues a message from the queue if any are available, or
//! waits for a message to be sent if the queue is empty. This module provides
//! an implementation of multi-producer, single-consumer channels built on top
//! of the lock-free queue implemented by [`ThingBuf`].
//!
//! The channel API in this module is broadly similar to the one provided by
//! other implementations, such as [`std::sync::mpsc`], [`tokio::sync::mpsc`],
//! or [`crossbeam::channel`]. A given channel instance is represented by a pair
//! of types:
//!
//! * a [`Sender`] handle, which represents the capacity to [send] messages to the
//!   channel
//! * a [`Receiver`] handle, which represents the capacity to [receive] messages
//!   from the channel
//!
//! ```
//! // Constructing a channel returns a paired `Sender` and `Receiver`:
//! let (tx, rx) = thingbuf::mpsc::channel::<String>(10);
//! # drop((tx, rx));
//! ```
//!
//! As these are multi-producer, single-consumer channels, the [`Sender`] type
//! implements `Clone`; it may be cloned any number of times to create multiple
//! [`Sender`]s that send messages to the same channel. On the other hand, each
//! channel instance has only a single [`Receiver`].
//!
//! # Disconnection
//!
//! When all [`Sender`] handles have been dropped, it is no longer
//! possible to send values into the channel. When this occurs, the channel is
//! considered to have been closed by the sender; calling [`Receiver::recv`]
//! once a channel has closed will return `None`. Note that if the [`Receiver`] has
//! not received any messages sent prior to the last [`Sender`] being dropped,
//! those messages will be returned before [`Receiver::recv`] returns `None`.
//!
//! If the [`Receiver`] handle is dropped, then messages can no longer
//! be read out of the channel. In this case, all further attempts to send will
//! result in an error.
//!
//! # Channel Flavors
//!
//! This module contains several different "flavors" of multi-producer,
//! single-consumer channels. The primary differences between the different
//! channel implementations are whether they wait asynchronously or by blocking
//! the current thread, and whether the array that stores channel messages is
//! allocated dynamically on the heap or statically at compile-time.
//!
//! |              | **Dynamic Allocation** | **Static Allocation**        |
//! |--------------|------------------------|------------------------------|
//! | **Async**    | [`channel`]            | [`StaticChannel`]            |
//! | **Blocking** | [`blocking::channel`]  | [`blocking::StaticChannel`]  |
//!
//! ## Asynchronous and Blocking Channels
//!
//! The default channel implementation (in this module) are _asynchronous_.
//! With these channels, the send and receive operations are [`async fn`]s
//! &mdash; when it's necessary to wait to send or receive a message, the
//! [`Future`]s returned by these functions will _yield_, allowing other [tasks]
//! to execute. These asynchronous channels are suitable for use in an async
//! runtime like [`tokio`], or in embedded or bare-metal systems that use
//! `async`/`await` syntax. They do not require the Rust standard library.
//!
//! If the "std" feature flag is enabled, this module also provides a
//! [`blocking`] submodule, which implements _blocking_ (or _synchronous_)
//! channels. The [blocking receiver] will instead wait for new messages by
//! blocking the  current thread, and the [blocking sender] will wait for send
//! capacity by blocking the current thread. A blocking channel can be
//! constructed using the [`blocking::channel`] function. Naturally, blocking
//! the current thread requires thread APIs, so these channels are only
//! available when the Rust standard library is available.
//!
//! An asynchronous channel is used with asynchronous tasks:
//!
//! ```rust
//! use thingbuf::mpsc;
//!
//! #[tokio::main]
//! async fn main() {
//!     let (tx, rx) = mpsc::channel(8);
//!
//!     // Spawn some tasks that write to the channel:
//!     for i in 0..10 {
//!         let tx = tx.clone();
//!         tokio::spawn(async move {
//!             tx.send(i).await.unwrap();
//!         });
//!     }
//!
//!     // Print out every message recieved from the channel:
//!     for _ in 0..10 {
//!         let j = rx.recv().await.unwrap();
//!
//!         println!("received {}", j);
//!         assert!(0 <= j && j < 10);
//!     }
//! }
//! ```
//!
//! A blocking channel is used with threads:
//!
//! ```rust
//! use thingbuf::mpsc::blocking;
//! use std::thread;
//!
//! let (tx, rx) = blocking::channel(8);
//!
//! // Spawn some threads that write to the channel:
//! for i in 0..10 {
//!     let tx = tx.clone();
//!     thread::spawn(move || {
//!         tx.send(i).unwrap();
//!     });
//! }
//!
//! // Print out every message recieved from the channel:
//! for _ in 0..10 {
//!     let j = rx.recv().unwrap();
//!
//!     println!("received {}", j);
//!     assert!(0 <= j && j < 10);
//! }
//! ```
//!
//! ## Static and Dynamically Allocated Channels
//!
//! The other difference between channel flavors is whether the array that backs
//! the channel's queue is allocated dynamically on the heap, or allocated
//! statically.
//!
//! The default channels returned by [`channel`] and [`blocking::channel`] are
//! _dynamically allocated_: although they are fixed-size, the size of the
//! channel can be determined at runtime when the channel is created, and any
//! number of dynamically-allocated channels may be created and destroyed over
//! the lifetime of the program. Because these channels dynamically allocate
//! memory, they require the "alloc" feature flag.
//!
//! In some use cases, though, dynamic memory allocation may not be available
//! (such as in some embedded systems, or within a memory allocator
//! implementation), or it may be desirable to avoid dynamic memory allocation
//! (such as in very performance-sensitive applications). To support those use
//! cases, `thingbuf::mpsc` also provides _static channels_. The size of these
//! channels is determined at compile-time (by a const generic parameter) rather
//! than at runtime, so they may be constructed in a static initializer. This
//! allows allocating the array that backs the channel's queue statically, so
//! that dynamic allocation is not needed.The [`StaticChannel`] and
//! [`blocking::StaticChannel`] types are used to construct static channels.
//!
//! A dynamically allocated channel's size can be determined at runtime:
//!
//! ```
//! use thingbuf::mpsc::blocking;
//! use std::env;
//!
//! # // the `main` fn is used explicitly to show that this is an "application".
//! # #[allow(clippy::needless_doctest_main)]
//! fn main() {
//!     // Determine the channel capacity from the first command-line line
//!     // argument, if one is present.
//!     let channel_capacity = env::args()
//!         .nth(1)
//!         .and_then(|cap| cap.parse::<usize>().ok())
//!         .unwrap_or(16);
//!
//!     // Construct a dynamically-sized blocking channel.
//!     let (tx, rx) = blocking::channel(channel_capacity);
//!
//!     tx.send("hello world").unwrap();
//!
//!     // ...
//!     # drop(tx); drop(rx);
//! }
//! ```
//!
//! A statically allocated channel may be used without heap allocation:
//!
//! ```
//! // We are in a `no_std` context with no memory allocator!
//! #![no_std]
//! use thingbuf::mpsc;
//!
//! // Create a channel backed by a static array with 256 entries.
//! static KERNEL_EVENTS: mpsc::StaticChannel<KernelEvent, 256> = mpsc::StaticChannel::new();
//!
//! #[no_mangle]
//! pub fn kernel_main() {
//!     // Split the static channel into a sender/receiver pair.
//!     let (event_tx, event_rx) = KERNEL_EVENTS.split();
//!     let mut kernel_tasks = TaskExecutor::new();
//!
//!     kernel_tasks.spawn(async move {
//!         // Process kernel events
//!         while let Some(event) = event_rx.recv().await {
//!             // ...
//!             # drop(event);
//!         }
//!     });
//!
//!     // Some device driver that needs to emit events to the channel.
//!     kernel_tasks.spawn(some_device_driver(event_tx.clone()));
//!
//!     loop {
//!         kernel_tasks.tick();
//!     }
//! }
//!
//! async fn some_device_driver(event_tx: mpsc::StaticSender<KernelEvent>) {
//!     let mut device = SomeDevice::new(0x42);
//!
//!     loop {
//!         // When the device has data, emit a kernel event.
//!         match device.poll_for_events().await {
//!             Ok(event) => event_tx.send(event).await.unwrap(),
//!             Err(err) => event_tx.send(KernelEvent::from(err)).await.unwrap(),
//!         }
//!     }
//! }
//! # type KernelEvent = ();
//! # struct TaskExecutor;
//! # impl TaskExecutor {
//! #   const fn new() -> Self { Self }
//! #   fn spawn(&mut self, _: impl core::future::Future) {}
//! #   fn tick(&mut self) {}
//! # }
//! # struct SomeDevice(u64);
//! # impl SomeDevice {
//! #   fn new(u: u64) -> Self { Self(u) }
//! #   async fn poll_for_events(&mut self) -> Result<(), ()> { Ok(()) }
//! # }
//! # fn main() {}
//! ```
//!
//! [send]: Sender::send
//! [receive]: Receiver::recv
//! [`ThingBuf`]: crate::ThingBuf
//! [`tokio::sync::mpsc`]: https://docs.rs/tokio/latest/tokio/sync/mpsc/index.html
//! [`crossbeam::channel`]: https://docs.rs/crossbeam/latest/crossbeam/channel/index.html
//! [`async fn`]: https://rust-lang.github.io/async-book/01_getting_started/04_async_await_primer.html
//! [`Future`]: core::future::Future
//! [tasks]: core::task
//! [`tokio`]: https://tokio.rs
//! [blocking receiver]: blocking::Receiver
//! [blocking sender]: blocking::Sender
use crate::{
    loom::{atomic::AtomicUsize, hint},
    recycling::{take, Recycle},
    wait::{Notify, WaitCell, WaitQueue, WaitResult},
    Core, Ref, Slot,
};
use core::{fmt, task::Poll};

pub mod errors;
use self::errors::{TryRecvError, TrySendError};

#[derive(Debug)]
struct ChannelCore<N> {
    core: Core,
    rx_wait: WaitCell<N>,
    tx_count: AtomicUsize,
    tx_wait: WaitQueue<N>,
}

struct SendRefInner<'a, T, N: Notify> {
    // /!\ LOAD BEARING STRUCT DROP ORDER /!\
    //
    // The `Ref` field *must* be dropped before the `NotifyInner` field, or else
    // loom tests will fail. This ensures that the mutable access to the slot is
    // considered to have ended *before* the receiver thread/task is notified.
    //
    // The alternatives to a load-bearing drop order would be:
    // (a) put one field inside an `Option` so it can be dropped before the
    //     other (not great, as it adds a little extra overhead even outside
    //     of Loom tests),
    // (b) use `core::mem::ManuallyDrop` (also not great, requires additional
    //     unsafe code that in this case we can avoid)
    //
    // So, given that, relying on struct field drop order seemed like the least
    // bad option here. Just don't reorder these fields. :)
    slot: Ref<'a, T>,
    _notify: NotifyRx<'a, N>,
}

struct RecvRefInner<'a, T, N: Notify + Unpin> {
    // /!\ LOAD BEARING STRUCT DROP ORDER /!\
    //
    // The `Ref` field *must* be dropped before the `NotifyTx` field, or else
    // loom tests will fail. This ensures that the mutable access to the slot is
    // considered to have ended *before* the receiver thread/task is notified.
    //
    // The alternatives to a load-bearing drop order would be:
    // (a) put one field inside an `Option` so it can be dropped before the
    //     other (not great, as it adds a little extra overhead even outside
    //     of Loom tests),
    // (b) use `core::mem::ManuallyDrop` (also not great, requires additional
    //     unsafe code that in this case we can avoid)
    //
    // So, given that, relying on struct field drop order seemed like the least
    // bad option here. Just don't reorder these fields. :)
    slot: Ref<'a, T>,
    _notify: crate::mpsc::NotifyTx<'a, N>,
}

struct NotifyRx<'a, N: Notify>(&'a WaitCell<N>);
struct NotifyTx<'a, N: Notify + Unpin>(&'a WaitQueue<N>);

// ==== impl Inner ====

impl<N> ChannelCore<N> {
    #[cfg(not(loom))]
    const fn new(capacity: usize) -> Self {
        Self {
            core: Core::new(capacity),
            rx_wait: WaitCell::new(),
            tx_count: AtomicUsize::new(1),
            tx_wait: WaitQueue::new(),
        }
    }

    #[cfg(loom)]
    fn new(capacity: usize) -> Self {
        Self {
            core: Core::new(capacity),
            rx_wait: WaitCell::new(),
            tx_count: AtomicUsize::new(1),
            tx_wait: WaitQueue::new(),
        }
    }
}

impl<N> ChannelCore<N>
where
    N: Notify + Unpin,
{
    fn close_rx(&self) {
        if self.core.close() {
            crate::loom::hint::spin_loop();
            test_println!("draining_queue");
            self.tx_wait.close();
        }
    }
}

impl<N> ChannelCore<N>
where
    N: Notify + Unpin,
{
    fn try_send_ref<'a, T, R>(
        &'a self,
        slots: &'a [Slot<T>],
        recycle: &R,
    ) -> Result<SendRefInner<'a, T, N>, TrySendError>
    where
        R: Recycle<T>,
    {
        self.core.push_ref(slots, recycle).map(|slot| SendRefInner {
            _notify: NotifyRx(&self.rx_wait),
            slot,
        })
    }

    fn try_send<T, R>(&self, slots: &[Slot<T>], val: T, recycle: &R) -> Result<(), TrySendError<T>>
    where
        R: Recycle<T>,
    {
        match self.try_send_ref(slots, recycle) {
            Ok(mut slot) => {
                slot.with_mut(|slot| *slot = val);
                Ok(())
            }
            Err(e) => Err(e.with_value(val)),
        }
    }

    fn try_recv_ref<'a, T>(
        &'a self,
        slots: &'a [Slot<T>],
    ) -> Result<RecvRefInner<'a, T, N>, TryRecvError> {
        self.core.pop_ref(slots).map(|slot| RecvRefInner {
            _notify: NotifyTx(&self.tx_wait),
            slot,
        })
    }

    fn try_recv<T, R>(&self, slots: &[Slot<T>], recycle: &R) -> Result<T, TryRecvError>
    where
        R: Recycle<T>,
    {
        match self.try_recv_ref(slots) {
            Ok(mut slot) => Ok(take(&mut *slot, recycle)),
            Err(e) => Err(e),
        }
    }

    /// Performs one iteration of the `recv_ref` loop.
    ///
    /// The loop itself has to be written in the actual `send` method's
    /// implementation, rather than on `inner`, because it might be async and
    /// may yield, or might park the thread.
    fn poll_recv_ref<'a, T>(
        &'a self,
        slots: &'a [Slot<T>],
        mk_waiter: impl Fn() -> N,
    ) -> Poll<Option<Ref<'a, T>>> {
        macro_rules! try_poll_recv {
            () => {
                // If we got a value, return it!
                match self.core.pop_ref(slots) {
                    Ok(slot) => return Poll::Ready(Some(slot)),
                    Err(TryRecvError::Closed) => return Poll::Ready(None),
                    _ => {}
                }
            };
        }

        test_println!("poll_recv_ref");
        loop {
            test_println!("poll_recv_ref => loop");

            // try to receive a reference, returning if we succeeded or the
            // channel is closed.
            try_poll_recv!();

            // otherwise, gotosleep
            match test_dbg!(self.rx_wait.wait_with(&mk_waiter)) {
                WaitResult::Wait => {
                    // we successfully registered a waiter! try polling again,
                    // just in case someone sent a message while we were
                    // registering the waiter.
                    try_poll_recv!();
                    test_println!("-> yield");
                    return Poll::Pending;
                }
                WaitResult::Closed => {
                    // the channel is closed (all the receivers are dropped).
                    // however, there may be messages left in the queue. try
                    // popping from the queue until it's empty.
                    return Poll::Ready(self.core.pop_ref(slots).ok());
                }
                WaitResult::Notified => {
                    // we were notified while we were trying to register the
                    // waiter. loop and try polling again.
                    hint::spin_loop();
                }
            }
        }
    }
}

// === impl SendRefInner ===

impl<T, N: Notify> core::ops::Deref for SendRefInner<'_, T, N> {
    type Target = T;

    #[inline]
    fn deref(&self) -> &Self::Target {
        self.slot.deref()
    }
}

impl<T, N: Notify> core::ops::DerefMut for SendRefInner<'_, T, N> {
    #[inline]
    fn deref_mut(&mut self) -> &mut Self::Target {
        self.slot.deref_mut()
    }
}

impl<T, N: Notify> SendRefInner<'_, T, N> {
    #[inline]
    pub(crate) fn with<U>(&self, f: impl FnOnce(&T) -> U) -> U {
        self.slot.with(f)
    }

    #[inline]
    pub(crate) fn with_mut<U>(&mut self, f: impl FnOnce(&mut T) -> U) -> U {
        self.slot.with_mut(f)
    }
}

impl<T: fmt::Debug, N: Notify> fmt::Debug for SendRefInner<'_, T, N> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.with(|val| fmt::Debug::fmt(val, f))
    }
}

impl<T: fmt::Display, N: Notify> fmt::Display for SendRefInner<'_, T, N> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.with(|val| fmt::Display::fmt(val, f))
    }
}

impl<T: fmt::Write, N: Notify> fmt::Write for SendRefInner<'_, T, N> {
    #[inline]
    fn write_str(&mut self, s: &str) -> fmt::Result {
        self.with_mut(|val| val.write_str(s))
    }

    #[inline]
    fn write_char(&mut self, c: char) -> fmt::Result {
        self.with_mut(|val| val.write_char(c))
    }

    #[inline]
    fn write_fmt(&mut self, f: fmt::Arguments<'_>) -> fmt::Result {
        self.with_mut(|val| val.write_fmt(f))
    }
}

// === impl RecvRefInner ===

impl<T, N: Notify + Unpin> core::ops::Deref for RecvRefInner<'_, T, N> {
    type Target = T;
    #[inline]
    fn deref(&self) -> &Self::Target {
        self.slot.deref()
    }
}

impl<T, N: Notify + Unpin> core::ops::DerefMut for RecvRefInner<'_, T, N> {
    #[inline]
    fn deref_mut(&mut self) -> &mut Self::Target {
        self.slot.deref_mut()
    }
}

impl<T: fmt::Debug, N: Notify + Unpin> fmt::Debug for RecvRefInner<'_, T, N> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.slot.fmt(f)
    }
}

impl<T: fmt::Display, N: Notify + Unpin> fmt::Display for RecvRefInner<'_, T, N> {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.slot.fmt(f)
    }
}

impl<T: fmt::Write, N: Notify + Unpin> fmt::Write for RecvRefInner<'_, T, N> {
    #[inline]
    fn write_str(&mut self, s: &str) -> fmt::Result {
        self.slot.write_str(s)
    }

    #[inline]
    fn write_char(&mut self, c: char) -> fmt::Result {
        self.slot.write_char(c)
    }

    #[inline]
    fn write_fmt(&mut self, f: fmt::Arguments<'_>) -> fmt::Result {
        self.slot.write_fmt(f)
    }
}

impl<N: Notify> Drop for NotifyRx<'_, N> {
    #[inline]
    fn drop(&mut self) {
        test_println!("notifying rx ({})", core::any::type_name::<N>());
        self.0.notify();
    }
}

impl<N: Notify + Unpin> Drop for NotifyTx<'_, N> {
    #[inline]
    fn drop(&mut self) {
        test_println!("notifying tx ({})", core::any::type_name::<N>());
        self.0.notify();
    }
}

macro_rules! impl_ref_inner {
    ($(#[$m:meta])*, $inner:ident, $name:ident, $notify:ty) => {
        $(#[$m])*
        pub struct $name<'a, T>($inner<'a, T, $notify>);

        impl<T> core::ops::Deref for $name<'_, T> {
            type Target = T;

            #[inline]
            fn deref(&self) -> &Self::Target {
                self.0.deref()
            }
        }

        impl<T> core::ops::DerefMut for $name<'_, T> {
            #[inline]
            fn deref_mut(&mut self) -> &mut Self::Target {
                self.0.deref_mut()
            }
        }

        impl<T: fmt::Debug> fmt::Debug for $name<'_, T> {
            fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
                self.0.fmt(f)
            }
        }

        impl<T: fmt::Display> fmt::Display for $name<'_, T> {
            fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
                self.0.fmt(f)
            }
        }

        impl<T: fmt::Write> fmt::Write for $name<'_, T> {
            #[inline]
            fn write_str(&mut self, s: &str) -> fmt::Result {
                self.0.write_str(s)
            }

            #[inline]
            fn write_char(&mut self, c: char) -> fmt::Result {
                self.0.write_char(c)
            }

            #[inline]
            fn write_fmt(&mut self, f: fmt::Arguments<'_>) -> fmt::Result {
                self.0.write_fmt(f)
            }
        }
    };
}

macro_rules! impl_send_ref {
    ($(#[$m:meta])* pub struct $name:ident<$notify:ty>;) => {
        impl_ref_inner!($(#[$m])*, SendRefInner, $name, $notify);
    };
}

macro_rules! impl_recv_ref {
    ($(#[$m:meta])* pub struct $name:ident<$notify:ty>;) => {
        impl_ref_inner!($(#[$m])*, RecvRefInner, $name, $notify);
    };
}

mod async_impl;
pub use self::async_impl::*;

feature! {
    #![feature = "std"]
    pub mod blocking;
}

#[cfg(all(loom, test))]
mod tests;