2013-12-06 02:56:17 +01:00
|
|
|
/* Copyright (c) 2010-2011 Dmitry Vyukov. All rights reserved.
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions are met:
|
|
|
|
*
|
|
|
|
* 1. Redistributions of source code must retain the above copyright notice,
|
|
|
|
* this list of conditions and the following disclaimer.
|
|
|
|
*
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY DMITRY VYUKOV "AS IS" AND ANY EXPRESS OR IMPLIED
|
|
|
|
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
|
|
|
|
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT
|
|
|
|
* SHALL DMITRY VYUKOV OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
|
|
|
|
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
|
|
|
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
|
|
|
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
|
|
|
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
|
|
|
|
* OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
|
|
|
|
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
*
|
|
|
|
* The views and conclusions contained in the software and documentation are
|
|
|
|
* those of the authors and should not be interpreted as representing official
|
|
|
|
* policies, either expressed or implied, of Dmitry Vyukov.
|
|
|
|
*/
|
|
|
|
|
|
|
|
// http://www.1024cores.net/home/lock-free-algorithms/queues/unbounded-spsc-queue
|
2013-12-13 02:27:37 +01:00
|
|
|
|
|
|
|
//! A single-producer single-consumer concurrent queue
|
|
|
|
//!
|
|
|
|
//! This module contains the implementation of an SPSC queue which can be used
|
|
|
|
//! concurrently between two tasks. This data structure is safe to use and
|
|
|
|
//! enforces the semantics that there is one pusher and one popper.
|
|
|
|
|
2015-01-23 03:22:03 +01:00
|
|
|
#![unstable(feature = "std_misc")]
|
2014-06-07 20:13:26 +02:00
|
|
|
|
|
|
|
use core::prelude::*;
|
|
|
|
|
2014-07-10 23:19:17 +02:00
|
|
|
use alloc::boxed::Box;
|
2014-06-07 20:13:26 +02:00
|
|
|
use core::mem;
|
2015-01-19 06:27:09 +01:00
|
|
|
use core::ptr;
|
std: Stabilize unit, bool, ty, tuple, arc, any
This commit applies stability attributes to the contents of these modules,
summarized here:
* The `unit` and `bool` modules have become #[unstable] as they are purely meant
for documentation purposes and are candidates for removal.
* The `ty` module has been deprecated, and the inner `Unsafe` type has been
renamed to `UnsafeCell` and moved to the `cell` module. The `marker1` field
has been removed as the compiler now always infers `UnsafeCell` to be
invariant. The `new` method i stable, but the `value` field, `get` and
`unwrap` methods are all unstable.
* The `tuple` module has its name as stable, the naming of the `TupleN` traits
as stable while the methods are all #[unstable]. The other impls in the module
have appropriate stability for the corresponding trait.
* The `arc` module has received the exact same treatment as the `rc` module
previously did.
* The `any` module has its name as stable. The `Any` trait is also stable, with
a new private supertrait which now contains the `get_type_id` method. This is
to make the method a private implementation detail rather than a public-facing
detail.
The two extension traits in the module are marked #[unstable] as they will not
be necessary with DST. The `is` method is #[stable], the as_{mut,ref} methods
have been renamed to downcast_{mut,ref} and are #[unstable].
The extension trait `BoxAny` has been clarified as to why it is unstable as it
will not be necessary with DST.
This is a breaking change because the `marker1` field was removed from the
`UnsafeCell` type. To deal with this change, you can simply delete the field and
only specify the value of the `data` field in static initializers.
[breaking-change]
2014-07-24 04:10:12 +02:00
|
|
|
use core::cell::UnsafeCell;
|
2014-06-07 20:13:26 +02:00
|
|
|
|
2015-01-10 22:42:48 +01:00
|
|
|
use sync::atomic::{AtomicPtr, AtomicUsize, Ordering};
|
2013-12-06 02:56:17 +01:00
|
|
|
|
|
|
|
// Node within the linked list queue of messages to send
|
|
|
|
struct Node<T> {
|
2014-01-26 09:43:42 +01:00
|
|
|
// FIXME: this could be an uninitialized T if we're careful enough, and
|
2013-12-06 02:56:17 +01:00
|
|
|
// that would reduce memory usage (and be a bit faster).
|
|
|
|
// is it worth it?
|
|
|
|
value: Option<T>, // nullable for re-use of nodes
|
|
|
|
next: AtomicPtr<Node<T>>, // next node in the queue
|
|
|
|
}
|
|
|
|
|
2014-01-07 00:23:37 +01:00
|
|
|
/// The single-producer single-consumer queue. This structure is not cloneable,
|
2014-05-20 02:50:57 +02:00
|
|
|
/// but it can be safely shared in an Arc if it is guaranteed that there
|
2014-01-07 00:23:37 +01:00
|
|
|
/// is only one popper and one pusher touching the queue at any one point in
|
|
|
|
/// time.
|
|
|
|
pub struct Queue<T> {
|
2013-12-06 02:56:17 +01:00
|
|
|
// consumer fields
|
std: Stabilize unit, bool, ty, tuple, arc, any
This commit applies stability attributes to the contents of these modules,
summarized here:
* The `unit` and `bool` modules have become #[unstable] as they are purely meant
for documentation purposes and are candidates for removal.
* The `ty` module has been deprecated, and the inner `Unsafe` type has been
renamed to `UnsafeCell` and moved to the `cell` module. The `marker1` field
has been removed as the compiler now always infers `UnsafeCell` to be
invariant. The `new` method i stable, but the `value` field, `get` and
`unwrap` methods are all unstable.
* The `tuple` module has its name as stable, the naming of the `TupleN` traits
as stable while the methods are all #[unstable]. The other impls in the module
have appropriate stability for the corresponding trait.
* The `arc` module has received the exact same treatment as the `rc` module
previously did.
* The `any` module has its name as stable. The `Any` trait is also stable, with
a new private supertrait which now contains the `get_type_id` method. This is
to make the method a private implementation detail rather than a public-facing
detail.
The two extension traits in the module are marked #[unstable] as they will not
be necessary with DST. The `is` method is #[stable], the as_{mut,ref} methods
have been renamed to downcast_{mut,ref} and are #[unstable].
The extension trait `BoxAny` has been clarified as to why it is unstable as it
will not be necessary with DST.
This is a breaking change because the `marker1` field was removed from the
`UnsafeCell` type. To deal with this change, you can simply delete the field and
only specify the value of the `data` field in static initializers.
[breaking-change]
2014-07-24 04:10:12 +02:00
|
|
|
tail: UnsafeCell<*mut Node<T>>, // where to pop from
|
2014-03-27 23:09:47 +01:00
|
|
|
tail_prev: AtomicPtr<Node<T>>, // where to pop from
|
2013-12-06 02:56:17 +01:00
|
|
|
|
|
|
|
// producer fields
|
std: Stabilize unit, bool, ty, tuple, arc, any
This commit applies stability attributes to the contents of these modules,
summarized here:
* The `unit` and `bool` modules have become #[unstable] as they are purely meant
for documentation purposes and are candidates for removal.
* The `ty` module has been deprecated, and the inner `Unsafe` type has been
renamed to `UnsafeCell` and moved to the `cell` module. The `marker1` field
has been removed as the compiler now always infers `UnsafeCell` to be
invariant. The `new` method i stable, but the `value` field, `get` and
`unwrap` methods are all unstable.
* The `tuple` module has its name as stable, the naming of the `TupleN` traits
as stable while the methods are all #[unstable]. The other impls in the module
have appropriate stability for the corresponding trait.
* The `arc` module has received the exact same treatment as the `rc` module
previously did.
* The `any` module has its name as stable. The `Any` trait is also stable, with
a new private supertrait which now contains the `get_type_id` method. This is
to make the method a private implementation detail rather than a public-facing
detail.
The two extension traits in the module are marked #[unstable] as they will not
be necessary with DST. The `is` method is #[stable], the as_{mut,ref} methods
have been renamed to downcast_{mut,ref} and are #[unstable].
The extension trait `BoxAny` has been clarified as to why it is unstable as it
will not be necessary with DST.
This is a breaking change because the `marker1` field was removed from the
`UnsafeCell` type. To deal with this change, you can simply delete the field and
only specify the value of the `data` field in static initializers.
[breaking-change]
2014-07-24 04:10:12 +02:00
|
|
|
head: UnsafeCell<*mut Node<T>>, // where to push to
|
|
|
|
first: UnsafeCell<*mut Node<T>>, // where to get new nodes from
|
|
|
|
tail_copy: UnsafeCell<*mut Node<T>>, // between first/tail
|
2013-12-06 02:56:17 +01:00
|
|
|
|
|
|
|
// Cache maintenance fields. Additions and subtractions are stored
|
|
|
|
// separately in order to allow them to use nonatomic addition/subtraction.
|
2014-03-27 23:09:47 +01:00
|
|
|
cache_bound: uint,
|
2015-01-10 22:42:48 +01:00
|
|
|
cache_additions: AtomicUsize,
|
|
|
|
cache_subtractions: AtomicUsize,
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
2014-12-22 00:49:42 +01:00
|
|
|
unsafe impl<T: Send> Send for Queue<T> { }
|
2014-12-06 17:39:25 +01:00
|
|
|
|
2014-12-22 00:49:42 +01:00
|
|
|
unsafe impl<T: Send> Sync for Queue<T> { }
|
2014-12-06 17:39:25 +01:00
|
|
|
|
2013-12-06 02:56:17 +01:00
|
|
|
impl<T: Send> Node<T> {
|
|
|
|
fn new() -> *mut Node<T> {
|
|
|
|
unsafe {
|
core: Remove the cast module
This commit revisits the `cast` module in libcore and libstd, and scrutinizes
all functions inside of it. The result was to remove the `cast` module entirely,
folding all functionality into the `mem` module. Specifically, this is the fate
of each function in the `cast` module.
* transmute - This function was moved to `mem`, but it is now marked as
#[unstable]. This is due to planned changes to the `transmute`
function and how it can be invoked (see the #[unstable] comment).
For more information, see RFC 5 and #12898
* transmute_copy - This function was moved to `mem`, with clarification that is
is not an error to invoke it with T/U that are different
sizes, but rather that it is strongly discouraged. This
function is now #[stable]
* forget - This function was moved to `mem` and marked #[stable]
* bump_box_refcount - This function was removed due to the deprecation of
managed boxes as well as its questionable utility.
* transmute_mut - This function was previously deprecated, and removed as part
of this commit.
* transmute_mut_unsafe - This function doesn't serve much of a purpose when it
can be achieved with an `as` in safe code, so it was
removed.
* transmute_lifetime - This function was removed because it is likely a strong
indication that code is incorrect in the first place.
* transmute_mut_lifetime - This function was removed for the same reasons as
`transmute_lifetime`
* copy_lifetime - This function was moved to `mem`, but it is marked
`#[unstable]` now due to the likelihood of being removed in
the future if it is found to not be very useful.
* copy_mut_lifetime - This function was also moved to `mem`, but had the same
treatment as `copy_lifetime`.
* copy_lifetime_vec - This function was removed because it is not used today,
and its existence is not necessary with DST
(copy_lifetime will suffice).
In summary, the cast module was stripped down to these functions, and then the
functions were moved to the `mem` module.
transmute - #[unstable]
transmute_copy - #[stable]
forget - #[stable]
copy_lifetime - #[unstable]
copy_mut_lifetime - #[unstable]
[breaking-change]
2014-05-09 19:34:51 +02:00
|
|
|
mem::transmute(box Node {
|
2013-12-06 02:56:17 +01:00
|
|
|
value: None,
|
2015-01-19 06:27:09 +01:00
|
|
|
next: AtomicPtr::new(ptr::null_mut::<Node<T>>()),
|
2013-12-06 02:56:17 +01:00
|
|
|
})
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-01-07 00:23:37 +01:00
|
|
|
impl<T: Send> Queue<T> {
|
2014-07-25 21:58:27 +02:00
|
|
|
/// Creates a new queue.
|
|
|
|
///
|
|
|
|
/// This is unsafe as the type system doesn't enforce a single
|
|
|
|
/// consumer-producer relationship. It also allows the consumer to `pop`
|
|
|
|
/// items while there is a `peek` active due to all methods having a
|
|
|
|
/// non-mutable receiver.
|
2014-01-07 00:23:37 +01:00
|
|
|
///
|
|
|
|
/// # Arguments
|
|
|
|
///
|
|
|
|
/// * `bound` - This queue implementation is implemented with a linked
|
|
|
|
/// list, and this means that a push is always a malloc. In
|
|
|
|
/// order to amortize this cost, an internal cache of nodes is
|
|
|
|
/// maintained to prevent a malloc from always being
|
|
|
|
/// necessary. This bound is the limit on the size of the
|
|
|
|
/// cache (if desired). If the value is 0, then the cache has
|
|
|
|
/// no bound. Otherwise, the cache will never grow larger than
|
|
|
|
/// `bound` (although the queue itself could be much larger.
|
2014-07-25 21:58:27 +02:00
|
|
|
pub unsafe fn new(bound: uint) -> Queue<T> {
|
2014-01-07 00:23:37 +01:00
|
|
|
let n1 = Node::new();
|
|
|
|
let n2 = Node::new();
|
2015-01-02 08:53:35 +01:00
|
|
|
(*n1).next.store(n2, Ordering::Relaxed);
|
2014-01-07 00:23:37 +01:00
|
|
|
Queue {
|
std: Stabilize unit, bool, ty, tuple, arc, any
This commit applies stability attributes to the contents of these modules,
summarized here:
* The `unit` and `bool` modules have become #[unstable] as they are purely meant
for documentation purposes and are candidates for removal.
* The `ty` module has been deprecated, and the inner `Unsafe` type has been
renamed to `UnsafeCell` and moved to the `cell` module. The `marker1` field
has been removed as the compiler now always infers `UnsafeCell` to be
invariant. The `new` method i stable, but the `value` field, `get` and
`unwrap` methods are all unstable.
* The `tuple` module has its name as stable, the naming of the `TupleN` traits
as stable while the methods are all #[unstable]. The other impls in the module
have appropriate stability for the corresponding trait.
* The `arc` module has received the exact same treatment as the `rc` module
previously did.
* The `any` module has its name as stable. The `Any` trait is also stable, with
a new private supertrait which now contains the `get_type_id` method. This is
to make the method a private implementation detail rather than a public-facing
detail.
The two extension traits in the module are marked #[unstable] as they will not
be necessary with DST. The `is` method is #[stable], the as_{mut,ref} methods
have been renamed to downcast_{mut,ref} and are #[unstable].
The extension trait `BoxAny` has been clarified as to why it is unstable as it
will not be necessary with DST.
This is a breaking change because the `marker1` field was removed from the
`UnsafeCell` type. To deal with this change, you can simply delete the field and
only specify the value of the `data` field in static initializers.
[breaking-change]
2014-07-24 04:10:12 +02:00
|
|
|
tail: UnsafeCell::new(n2),
|
2014-01-07 00:23:37 +01:00
|
|
|
tail_prev: AtomicPtr::new(n1),
|
std: Stabilize unit, bool, ty, tuple, arc, any
This commit applies stability attributes to the contents of these modules,
summarized here:
* The `unit` and `bool` modules have become #[unstable] as they are purely meant
for documentation purposes and are candidates for removal.
* The `ty` module has been deprecated, and the inner `Unsafe` type has been
renamed to `UnsafeCell` and moved to the `cell` module. The `marker1` field
has been removed as the compiler now always infers `UnsafeCell` to be
invariant. The `new` method i stable, but the `value` field, `get` and
`unwrap` methods are all unstable.
* The `tuple` module has its name as stable, the naming of the `TupleN` traits
as stable while the methods are all #[unstable]. The other impls in the module
have appropriate stability for the corresponding trait.
* The `arc` module has received the exact same treatment as the `rc` module
previously did.
* The `any` module has its name as stable. The `Any` trait is also stable, with
a new private supertrait which now contains the `get_type_id` method. This is
to make the method a private implementation detail rather than a public-facing
detail.
The two extension traits in the module are marked #[unstable] as they will not
be necessary with DST. The `is` method is #[stable], the as_{mut,ref} methods
have been renamed to downcast_{mut,ref} and are #[unstable].
The extension trait `BoxAny` has been clarified as to why it is unstable as it
will not be necessary with DST.
This is a breaking change because the `marker1` field was removed from the
`UnsafeCell` type. To deal with this change, you can simply delete the field and
only specify the value of the `data` field in static initializers.
[breaking-change]
2014-07-24 04:10:12 +02:00
|
|
|
head: UnsafeCell::new(n2),
|
|
|
|
first: UnsafeCell::new(n1),
|
|
|
|
tail_copy: UnsafeCell::new(n1),
|
2014-01-07 00:23:37 +01:00
|
|
|
cache_bound: bound,
|
2015-01-10 22:42:48 +01:00
|
|
|
cache_additions: AtomicUsize::new(0),
|
|
|
|
cache_subtractions: AtomicUsize::new(0),
|
2014-01-07 00:23:37 +01:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
2014-01-07 00:23:37 +01:00
|
|
|
/// Pushes a new value onto this queue. Note that to use this function
|
|
|
|
/// safely, it must be externally guaranteed that there is only one pusher.
|
2014-05-20 02:32:04 +02:00
|
|
|
pub fn push(&self, t: T) {
|
2014-01-07 00:23:37 +01:00
|
|
|
unsafe {
|
|
|
|
// Acquire a node (which either uses a cached one or allocates a new
|
|
|
|
// one), and then append this to the 'head' node.
|
|
|
|
let n = self.alloc();
|
|
|
|
assert!((*n).value.is_none());
|
|
|
|
(*n).value = Some(t);
|
2015-01-19 06:27:09 +01:00
|
|
|
(*n).next.store(ptr::null_mut(), Ordering::Relaxed);
|
2015-01-02 08:53:35 +01:00
|
|
|
(**self.head.get()).next.store(n, Ordering::Release);
|
2014-05-20 02:32:04 +02:00
|
|
|
*self.head.get() = n;
|
2014-01-07 00:23:37 +01:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
2014-05-20 02:32:04 +02:00
|
|
|
unsafe fn alloc(&self) -> *mut Node<T> {
|
2013-12-06 02:56:17 +01:00
|
|
|
// First try to see if we can consume the 'first' node for our uses.
|
|
|
|
// We try to avoid as many atomic instructions as possible here, so
|
|
|
|
// the addition to cache_subtractions is not atomic (plus we're the
|
|
|
|
// only one subtracting from the cache).
|
2014-05-20 02:32:04 +02:00
|
|
|
if *self.first.get() != *self.tail_copy.get() {
|
2013-12-06 02:56:17 +01:00
|
|
|
if self.cache_bound > 0 {
|
2015-01-02 08:53:35 +01:00
|
|
|
let b = self.cache_subtractions.load(Ordering::Relaxed);
|
|
|
|
self.cache_subtractions.store(b + 1, Ordering::Relaxed);
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
2014-05-20 02:32:04 +02:00
|
|
|
let ret = *self.first.get();
|
2015-01-02 08:53:35 +01:00
|
|
|
*self.first.get() = (*ret).next.load(Ordering::Relaxed);
|
2013-12-06 02:56:17 +01:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
// If the above fails, then update our copy of the tail and try
|
|
|
|
// again.
|
2015-01-02 08:53:35 +01:00
|
|
|
*self.tail_copy.get() = self.tail_prev.load(Ordering::Acquire);
|
2014-05-20 02:32:04 +02:00
|
|
|
if *self.first.get() != *self.tail_copy.get() {
|
2013-12-06 02:56:17 +01:00
|
|
|
if self.cache_bound > 0 {
|
2015-01-02 08:53:35 +01:00
|
|
|
let b = self.cache_subtractions.load(Ordering::Relaxed);
|
|
|
|
self.cache_subtractions.store(b + 1, Ordering::Relaxed);
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
2014-05-20 02:32:04 +02:00
|
|
|
let ret = *self.first.get();
|
2015-01-02 08:53:35 +01:00
|
|
|
*self.first.get() = (*ret).next.load(Ordering::Relaxed);
|
2013-12-06 02:56:17 +01:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
// If all of that fails, then we have to allocate a new node
|
|
|
|
// (there's nothing in the node cache).
|
|
|
|
Node::new()
|
|
|
|
}
|
|
|
|
|
2014-01-07 00:23:37 +01:00
|
|
|
/// Attempts to pop a value from this queue. Remember that to use this type
|
|
|
|
/// safely you must ensure that there is only one popper at a time.
|
2014-05-20 02:32:04 +02:00
|
|
|
pub fn pop(&self) -> Option<T> {
|
2014-01-07 00:23:37 +01:00
|
|
|
unsafe {
|
|
|
|
// The `tail` node is not actually a used node, but rather a
|
|
|
|
// sentinel from where we should start popping from. Hence, look at
|
|
|
|
// tail's next field and see if we can use it. If we do a pop, then
|
|
|
|
// the current tail node is a candidate for going into the cache.
|
2014-05-20 02:32:04 +02:00
|
|
|
let tail = *self.tail.get();
|
2015-01-02 08:53:35 +01:00
|
|
|
let next = (*tail).next.load(Ordering::Acquire);
|
2014-01-07 00:23:37 +01:00
|
|
|
if next.is_null() { return None }
|
|
|
|
assert!((*next).value.is_some());
|
|
|
|
let ret = (*next).value.take();
|
2013-12-06 02:56:17 +01:00
|
|
|
|
2014-05-20 02:32:04 +02:00
|
|
|
*self.tail.get() = next;
|
2014-01-07 00:23:37 +01:00
|
|
|
if self.cache_bound == 0 {
|
2015-01-02 08:53:35 +01:00
|
|
|
self.tail_prev.store(tail, Ordering::Release);
|
2013-12-06 02:56:17 +01:00
|
|
|
} else {
|
2014-01-07 00:23:37 +01:00
|
|
|
// FIXME: this is dubious with overflow.
|
2015-01-02 08:53:35 +01:00
|
|
|
let additions = self.cache_additions.load(Ordering::Relaxed);
|
|
|
|
let subtractions = self.cache_subtractions.load(Ordering::Relaxed);
|
2014-01-07 00:23:37 +01:00
|
|
|
let size = additions - subtractions;
|
|
|
|
|
|
|
|
if size < self.cache_bound {
|
2015-01-02 08:53:35 +01:00
|
|
|
self.tail_prev.store(tail, Ordering::Release);
|
|
|
|
self.cache_additions.store(additions + 1, Ordering::Relaxed);
|
2014-01-07 00:23:37 +01:00
|
|
|
} else {
|
2015-01-02 08:53:35 +01:00
|
|
|
(*self.tail_prev.load(Ordering::Relaxed))
|
|
|
|
.next.store(next, Ordering::Relaxed);
|
2014-01-07 00:23:37 +01:00
|
|
|
// We have successfully erased all references to 'tail', so
|
|
|
|
// now we can safely drop it.
|
core: Remove the cast module
This commit revisits the `cast` module in libcore and libstd, and scrutinizes
all functions inside of it. The result was to remove the `cast` module entirely,
folding all functionality into the `mem` module. Specifically, this is the fate
of each function in the `cast` module.
* transmute - This function was moved to `mem`, but it is now marked as
#[unstable]. This is due to planned changes to the `transmute`
function and how it can be invoked (see the #[unstable] comment).
For more information, see RFC 5 and #12898
* transmute_copy - This function was moved to `mem`, with clarification that is
is not an error to invoke it with T/U that are different
sizes, but rather that it is strongly discouraged. This
function is now #[stable]
* forget - This function was moved to `mem` and marked #[stable]
* bump_box_refcount - This function was removed due to the deprecation of
managed boxes as well as its questionable utility.
* transmute_mut - This function was previously deprecated, and removed as part
of this commit.
* transmute_mut_unsafe - This function doesn't serve much of a purpose when it
can be achieved with an `as` in safe code, so it was
removed.
* transmute_lifetime - This function was removed because it is likely a strong
indication that code is incorrect in the first place.
* transmute_mut_lifetime - This function was removed for the same reasons as
`transmute_lifetime`
* copy_lifetime - This function was moved to `mem`, but it is marked
`#[unstable]` now due to the likelihood of being removed in
the future if it is found to not be very useful.
* copy_mut_lifetime - This function was also moved to `mem`, but had the same
treatment as `copy_lifetime`.
* copy_lifetime_vec - This function was removed because it is not used today,
and its existence is not necessary with DST
(copy_lifetime will suffice).
In summary, the cast module was stripped down to these functions, and then the
functions were moved to the `mem` module.
transmute - #[unstable]
transmute_copy - #[stable]
forget - #[stable]
copy_lifetime - #[unstable]
copy_mut_lifetime - #[unstable]
[breaking-change]
2014-05-09 19:34:51 +02:00
|
|
|
let _: Box<Node<T>> = mem::transmute(tail);
|
2014-01-07 00:23:37 +01:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
2014-01-07 00:23:37 +01:00
|
|
|
return ret;
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Rewrite channels yet again for upgradeability
This, the Nth rewrite of channels, is not a rewrite of the core logic behind
channels, but rather their API usage. In the past, we had the distinction
between oneshot, stream, and shared channels, but the most recent rewrite
dropped oneshots in favor of streams and shared channels.
This distinction of stream vs shared has shown that it's not quite what we'd
like either, and this moves the `std::comm` module in the direction of "one
channel to rule them all". There now remains only one Chan and one Port.
This new channel is actually a hybrid oneshot/stream/shared channel under the
hood in order to optimize for the use cases in question. Additionally, this also
reduces the cognitive burden of having to choose between a Chan or a SharedChan
in an API.
My simple benchmarks show no reduction in efficiency over the existing channels
today, and a 3x improvement in the oneshot case. I sadly don't have a
pre-last-rewrite compiler to test out the old old oneshots, but I would imagine
that the performance is comparable, but slightly slower (due to atomic reference
counting).
This commit also brings the bonus bugfix to channels that the pending queue of
messages are all dropped when a Port disappears rather then when both the Port
and the Chan disappear.
2014-01-09 03:31:48 +01:00
|
|
|
/// Attempts to peek at the head of the queue, returning `None` if the queue
|
|
|
|
/// has no data currently
|
2014-07-25 21:58:27 +02:00
|
|
|
///
|
|
|
|
/// # Warning
|
|
|
|
/// The reference returned is invalid if it is not used before the consumer
|
|
|
|
/// pops the value off the queue. If the producer then pushes another value
|
|
|
|
/// onto the queue, it will overwrite the value pointed to by the reference.
|
2014-05-20 02:32:04 +02:00
|
|
|
pub fn peek<'a>(&'a self) -> Option<&'a mut T> {
|
Rewrite channels yet again for upgradeability
This, the Nth rewrite of channels, is not a rewrite of the core logic behind
channels, but rather their API usage. In the past, we had the distinction
between oneshot, stream, and shared channels, but the most recent rewrite
dropped oneshots in favor of streams and shared channels.
This distinction of stream vs shared has shown that it's not quite what we'd
like either, and this moves the `std::comm` module in the direction of "one
channel to rule them all". There now remains only one Chan and one Port.
This new channel is actually a hybrid oneshot/stream/shared channel under the
hood in order to optimize for the use cases in question. Additionally, this also
reduces the cognitive burden of having to choose between a Chan or a SharedChan
in an API.
My simple benchmarks show no reduction in efficiency over the existing channels
today, and a 3x improvement in the oneshot case. I sadly don't have a
pre-last-rewrite compiler to test out the old old oneshots, but I would imagine
that the performance is comparable, but slightly slower (due to atomic reference
counting).
This commit also brings the bonus bugfix to channels that the pending queue of
messages are all dropped when a Port disappears rather then when both the Port
and the Chan disappear.
2014-01-09 03:31:48 +01:00
|
|
|
// This is essentially the same as above with all the popping bits
|
|
|
|
// stripped out.
|
2014-01-07 00:23:37 +01:00
|
|
|
unsafe {
|
2014-05-20 02:32:04 +02:00
|
|
|
let tail = *self.tail.get();
|
2015-01-02 08:53:35 +01:00
|
|
|
let next = (*tail).next.load(Ordering::Acquire);
|
Rewrite channels yet again for upgradeability
This, the Nth rewrite of channels, is not a rewrite of the core logic behind
channels, but rather their API usage. In the past, we had the distinction
between oneshot, stream, and shared channels, but the most recent rewrite
dropped oneshots in favor of streams and shared channels.
This distinction of stream vs shared has shown that it's not quite what we'd
like either, and this moves the `std::comm` module in the direction of "one
channel to rule them all". There now remains only one Chan and one Port.
This new channel is actually a hybrid oneshot/stream/shared channel under the
hood in order to optimize for the use cases in question. Additionally, this also
reduces the cognitive burden of having to choose between a Chan or a SharedChan
in an API.
My simple benchmarks show no reduction in efficiency over the existing channels
today, and a 3x improvement in the oneshot case. I sadly don't have a
pre-last-rewrite compiler to test out the old old oneshots, but I would imagine
that the performance is comparable, but slightly slower (due to atomic reference
counting).
This commit also brings the bonus bugfix to channels that the pending queue of
messages are all dropped when a Port disappears rather then when both the Port
and the Chan disappear.
2014-01-09 03:31:48 +01:00
|
|
|
if next.is_null() { return None }
|
|
|
|
return (*next).value.as_mut();
|
2014-01-07 00:23:37 +01:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#[unsafe_destructor]
|
2014-01-07 00:23:37 +01:00
|
|
|
impl<T: Send> Drop for Queue<T> {
|
2013-12-06 02:56:17 +01:00
|
|
|
fn drop(&mut self) {
|
|
|
|
unsafe {
|
2014-05-20 02:32:04 +02:00
|
|
|
let mut cur = *self.first.get();
|
2013-12-06 02:56:17 +01:00
|
|
|
while !cur.is_null() {
|
2015-01-02 08:53:35 +01:00
|
|
|
let next = (*cur).next.load(Ordering::Relaxed);
|
core: Remove the cast module
This commit revisits the `cast` module in libcore and libstd, and scrutinizes
all functions inside of it. The result was to remove the `cast` module entirely,
folding all functionality into the `mem` module. Specifically, this is the fate
of each function in the `cast` module.
* transmute - This function was moved to `mem`, but it is now marked as
#[unstable]. This is due to planned changes to the `transmute`
function and how it can be invoked (see the #[unstable] comment).
For more information, see RFC 5 and #12898
* transmute_copy - This function was moved to `mem`, with clarification that is
is not an error to invoke it with T/U that are different
sizes, but rather that it is strongly discouraged. This
function is now #[stable]
* forget - This function was moved to `mem` and marked #[stable]
* bump_box_refcount - This function was removed due to the deprecation of
managed boxes as well as its questionable utility.
* transmute_mut - This function was previously deprecated, and removed as part
of this commit.
* transmute_mut_unsafe - This function doesn't serve much of a purpose when it
can be achieved with an `as` in safe code, so it was
removed.
* transmute_lifetime - This function was removed because it is likely a strong
indication that code is incorrect in the first place.
* transmute_mut_lifetime - This function was removed for the same reasons as
`transmute_lifetime`
* copy_lifetime - This function was moved to `mem`, but it is marked
`#[unstable]` now due to the likelihood of being removed in
the future if it is found to not be very useful.
* copy_mut_lifetime - This function was also moved to `mem`, but had the same
treatment as `copy_lifetime`.
* copy_lifetime_vec - This function was removed because it is not used today,
and its existence is not necessary with DST
(copy_lifetime will suffice).
In summary, the cast module was stripped down to these functions, and then the
functions were moved to the `mem` module.
transmute - #[unstable]
transmute_copy - #[stable]
forget - #[stable]
copy_lifetime - #[unstable]
copy_mut_lifetime - #[unstable]
[breaking-change]
2014-05-09 19:34:51 +02:00
|
|
|
let _n: Box<Node<T>> = mem::transmute(cur);
|
2013-12-06 02:56:17 +01:00
|
|
|
cur = next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#[cfg(test)]
|
|
|
|
mod test {
|
2014-12-22 18:04:23 +01:00
|
|
|
use prelude::v1::*;
|
2014-05-20 02:50:57 +02:00
|
|
|
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
use sync::Arc;
|
|
|
|
use super::Queue;
|
2014-12-22 18:04:23 +01:00
|
|
|
use thread::Thread;
|
2014-12-23 20:53:35 +01:00
|
|
|
use sync::mpsc::channel;
|
2013-12-06 02:56:17 +01:00
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn smoke() {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe {
|
|
|
|
let queue = Queue::new(0);
|
2015-01-25 22:05:03 +01:00
|
|
|
queue.push(1);
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
queue.push(2);
|
2015-01-25 22:05:03 +01:00
|
|
|
assert_eq!(queue.pop(), Some(1));
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
assert_eq!(queue.pop(), Some(2));
|
|
|
|
assert_eq!(queue.pop(), None);
|
|
|
|
queue.push(3);
|
|
|
|
queue.push(4);
|
|
|
|
assert_eq!(queue.pop(), Some(3));
|
|
|
|
assert_eq!(queue.pop(), Some(4));
|
|
|
|
assert_eq!(queue.pop(), None);
|
|
|
|
}
|
2014-07-25 21:58:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn peek() {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe {
|
|
|
|
let queue = Queue::new(0);
|
2015-01-25 22:05:03 +01:00
|
|
|
queue.push(vec![1]);
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
|
|
|
|
// Ensure the borrowchecker works
|
|
|
|
match queue.peek() {
|
|
|
|
Some(vec) => match vec.as_slice() {
|
|
|
|
// Note that `pop` is not allowed here due to borrow
|
|
|
|
[1] => {}
|
|
|
|
_ => return
|
|
|
|
},
|
|
|
|
None => unreachable!()
|
|
|
|
}
|
2014-07-25 21:58:27 +02:00
|
|
|
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
queue.pop();
|
2014-07-25 21:58:27 +02:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn drop_full() {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe {
|
|
|
|
let q = Queue::new(0);
|
2015-01-25 22:05:03 +01:00
|
|
|
q.push(box 1);
|
|
|
|
q.push(box 2);
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn smoke_bound() {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe {
|
|
|
|
let q = Queue::new(0);
|
2015-01-25 22:05:03 +01:00
|
|
|
q.push(1);
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
q.push(2);
|
|
|
|
assert_eq!(q.pop(), Some(1));
|
|
|
|
assert_eq!(q.pop(), Some(2));
|
|
|
|
assert_eq!(q.pop(), None);
|
|
|
|
q.push(3);
|
|
|
|
q.push(4);
|
|
|
|
assert_eq!(q.pop(), Some(3));
|
|
|
|
assert_eq!(q.pop(), Some(4));
|
|
|
|
assert_eq!(q.pop(), None);
|
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn stress() {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe {
|
|
|
|
stress_bound(0);
|
|
|
|
stress_bound(1);
|
|
|
|
}
|
2013-12-06 02:56:17 +01:00
|
|
|
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
unsafe fn stress_bound(bound: uint) {
|
|
|
|
let q = Arc::new(Queue::new(bound));
|
2014-07-25 21:58:27 +02:00
|
|
|
|
2014-03-09 22:58:32 +01:00
|
|
|
let (tx, rx) = channel();
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
let q2 = q.clone();
|
2014-12-22 18:04:23 +01:00
|
|
|
let _t = Thread::spawn(move|| {
|
2015-01-26 21:46:12 +01:00
|
|
|
for _ in 0u..100000 {
|
2013-12-06 02:56:17 +01:00
|
|
|
loop {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
match q2.pop() {
|
2015-01-25 22:05:03 +01:00
|
|
|
Some(1) => break,
|
2014-10-09 21:17:22 +02:00
|
|
|
Some(_) => panic!(),
|
2013-12-06 02:56:17 +01:00
|
|
|
None => {}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2014-12-23 20:53:35 +01:00
|
|
|
tx.send(()).unwrap();
|
2014-01-27 04:42:26 +01:00
|
|
|
});
|
2015-01-25 22:05:03 +01:00
|
|
|
for _ in 0..100000 {
|
std: Rewrite the `sync` module
This commit is a reimplementation of `std::sync` to be based on the
system-provided primitives wherever possible. The previous implementation was
fundamentally built on top of channels, and as part of the runtime reform it has
become clear that this is not the level of abstraction that the standard level
should be providing. This rewrite aims to provide as thin of a shim as possible
on top of the system primitives in order to make them safe.
The overall interface of the `std::sync` module has in general not changed, but
there are a few important distinctions, highlighted below:
* The condition variable type, `Condvar`, has been separated out of a `Mutex`.
A condition variable is now an entirely separate type. This separation
benefits users who only use one mutex, and provides a clearer distinction of
who's responsible for managing condition variables (the application).
* All of `Condvar`, `Mutex`, and `RWLock` are now directly built on top of
system primitives rather than using a custom implementation. The `Once`,
`Barrier`, and `Semaphore` types are still built upon these abstractions of
the system primitives.
* The `Condvar`, `Mutex`, and `RWLock` types all have a new static type and
constant initializer corresponding to them. These are provided primarily for C
FFI interoperation, but are often useful to otherwise simply have a global
lock. The types, however, will leak memory unless `destroy()` is called on
them, which is clearly documented.
* The `Condvar` implementation for an `RWLock` write lock has been removed. This
may be added back in the future with a userspace implementation, but this
commit is focused on exposing the system primitives first.
* The fundamental architecture of this design is to provide two separate layers.
The first layer is that exposed by `sys_common` which is a cross-platform
bare-metal abstraction of the system synchronization primitives. No attempt is
made at making this layer safe, and it is quite unsafe to use! It is currently
not exported as part of the API of the standard library, but the stabilization
of the `sys` module will ensure that these will be exposed in time. The
purpose of this layer is to provide the core cross-platform abstractions if
necessary to implementors.
The second layer is the layer provided by `std::sync` which is intended to be
the thinnest possible layer on top of `sys_common` which is entirely safe to
use. There are a few concerns which need to be addressed when making these
system primitives safe:
* Once used, the OS primitives can never be **moved**. This means that they
essentially need to have a stable address. The static primitives use
`&'static self` to enforce this, and the non-static primitives all use a
`Box` to provide this guarantee.
* Poisoning is leveraged to ensure that invalid data is not accessible from
other tasks after one has panicked.
In addition to these overall blanket safety limitations, each primitive has a
few restrictions of its own:
* Mutexes and rwlocks can only be unlocked from the same thread that they
were locked by. This is achieved through RAII lock guards which cannot be
sent across threads.
* Mutexes and rwlocks can only be unlocked if they were previously locked.
This is achieved by not exposing an unlocking method.
* A condition variable can only be waited on with a locked mutex. This is
achieved by requiring a `MutexGuard` in the `wait()` method.
* A condition variable cannot be used concurrently with more than one mutex.
This is guaranteed by dynamically binding a condition variable to
precisely one mutex for its entire lifecycle. This restriction may be able
to be relaxed in the future (a mutex is unbound when no threads are
waiting on the condvar), but for now it is sufficient to guarantee safety.
* Condvars now support timeouts for their blocking operations. The
implementation for these operations is provided by the system.
Due to the modification of the `Condvar` API, removal of the `std::sync::mutex`
API, and reimplementation, this is a breaking change. Most code should be fairly
easy to port using the examples in the documentation of these primitives.
[breaking-change]
Closes #17094
Closes #18003
2014-11-24 20:16:40 +01:00
|
|
|
q.push(1);
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
2014-12-23 20:53:35 +01:00
|
|
|
rx.recv().unwrap();
|
2013-12-06 02:56:17 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|