Struct crossbeam_epoch::Guard

pub struct Guard { /* fields omitted */ }

A guard that keeps the current thread pinned.

Pinning

The current thread is pinned by calling pin, which returns a new guard:

use crossbeam_epoch as epoch;

// It is often convenient to prefix a call to `pin` with a `&` in order to create a reference.
// This is not really necessary, but makes passing references to the guard a bit easier.
let guard = &epoch::pin();

When a guard gets dropped, the current thread is automatically unpinned.

Pointers on the stack

Having a guard allows us to create pointers on the stack to heap-allocated objects. For example:

use crossbeam_epoch::{self as epoch, Atomic, Owned};
use std::sync::atomic::Ordering::SeqCst;

// Create a heap-allocated number.
let a = Atomic::new(777);

// Pin the current thread.
let guard = &epoch::pin();

// Load the heap-allocated object and create pointer `p` on the stack.
let p = a.load(SeqCst, guard);

// Dereference the pointer and print the value:
if let Some(num) = unsafe { p.as_ref() } {
    println!("The number is {}.", num);
}

Multiple guards

Pinning is reentrant and it is perfectly legal to create multiple guards. In that case, the thread will actually be pinned only when the first guard is created and unpinned when the last one is dropped:

use crossbeam_epoch as epoch;

let guard1 = epoch::pin();
let guard2 = epoch::pin();
assert!(epoch::is_pinned());
drop(guard1);
assert!(epoch::is_pinned());
drop(guard2);
assert!(!epoch::is_pinned());

The same can be achieved by cloning guards:

use crossbeam_epoch as epoch;

let guard1 = epoch::pin();
let guard2 = guard1.clone();

Methods

impl Guard

unsafe fn defer<F, R>(&self, f: F) where
    F: FnOnce() -> R, 

Stores a function so that it can be executed at some point after all currently pinned threads get unpinned.

This method first stores f into the thread-local (or handle-local) cache. If this cache becomes full, some functions are moved into the global cache. At the same time, some functions from both local and global caches may get executed in order to incrementally clean up the caches as they fill up.

There is no guarantee when exactly f will be executed. The only guarantee is that won't until all currently pinned threads get unpinned. In theory, f might never be deallocated, but the epoch-based garbage collection will make an effort to execute it reasonably soon.

If this method is called from an unprotected guard, the function will simply be executed immediately.

Safety

The given function must not hold reference onto the stack. It is highly recommended that the passed function is always marked with move in order to prevent accidental borrows.

use crossbeam_epoch as epoch;

let guard = &epoch::pin();
let message = "Hello!";
unsafe {
    // ALWAYS use `move` when sending a closure into `defef`.
    guard.defer(move || {
        println!("{}", message);
    });
}

Apart from that, keep in mind that another thread may execute f, so anything accessed by the closure must be Send.

Examples

When a heap-allocated object in a data structure becomes unreachable, it has to be deallocated. However, the current thread and other threads may be still holding references on the stack to that same object. Therefore it cannot be deallocated before those references get dropped. This method can defer deallocation until all those threads get unpinned and consequently drop all their references on the stack.

use crossbeam_epoch::{self as epoch, Atomic, Owned};
use std::sync::atomic::Ordering::SeqCst;

let a = Atomic::new("foo");

// Now suppose that `a` is shared among multiple threads and concurrently
// accessed and modified...

// Pin the current thread.
let guard = &epoch::pin();

// Steal the object currently stored in `a` and swap it with another one.
let p = a.swap(Owned::new("bar").into_shared(guard), SeqCst, guard);

if !p.is_null() {
    // The object `p` is pointing to is now unreachable.
    // Defer its deallocation until all currently pinned threads get unpinned.
    unsafe {
        // ALWAYS use `move` when sending a closure into `defer`.
        guard.defer(move || {
            println!("{} is now being deallocated.", p.deref());
            // Now we have unique access to the object pointed to by `p` and can turn it
            // into an `Owned`. Dropping the `Owned` will deallocate the object.
            drop(p.into_owned());
        });
    }
}

fn flush(&self)

Clears up the thread-local cache of deferred functions by executing them or moving into the global cache.

Call this method after deferring execution of a function if you want to get it executed as soon as possible. Flushing will make sure it is residing in in the global cache, so that any thread has a chance of taking the function and executing it.

If this method is called from an unprotected guard, it is a no-op (nothing happens).

Examples

use crossbeam_epoch as epoch;

let guard = &epoch::pin();
unsafe {
    guard.defer(move || {
        println!("This better be printed as soon as possible!");
    });
}
guard.flush();

fn repin(&mut self)

Unpins and then immediately re-pins the thread.

This method is useful when you don't want delay the advancement of the global epoch by holding an old epoch. For safety, you should not maintain any guard-based reference across the call (the latter is enforced by &mut self). The thread will only be repinned if this is the only active guard for the current thread.

If this method is called from an unprotected guard, then the call will be just no-op.

Examples

use crossbeam_epoch::{self as epoch, Atomic};
use std::sync::atomic::Ordering::SeqCst;
use std::thread;
use std::time::Duration;

let a = Atomic::new(777);
let mut guard = epoch::pin();
{
    let p = a.load(SeqCst, &guard);
    assert_eq!(unsafe { p.as_ref() }, Some(&777));
}
guard.repin();
{
    let p = a.load(SeqCst, &guard);
    assert_eq!(unsafe { p.as_ref() }, Some(&777));
}

fn repin_after<F, R>(&mut self, f: F) -> R where
    F: FnOnce() -> R, 

Temporarily unpins the thread, executes the given function and then re-pins the thread.

This method is useful when you need to perform a long-running operation (e.g. sleeping) and don't need to maintain any guard-based reference across the call (the latter is enforced by &mut self). The thread will only be unpinned if this is the only active guard for the current thread.

If this method is called from an unprotected guard, then the passed function is called directly without unpinning the thread.

Examples

use crossbeam_epoch::{self as epoch, Atomic};
use std::sync::atomic::Ordering::SeqCst;
use std::thread;
use std::time::Duration;

let a = Atomic::new(777);
let mut guard = epoch::pin();
{
    let p = a.load(SeqCst, &guard);
    assert_eq!(unsafe { p.as_ref() }, Some(&777));
}
guard.repin_after(|| thread::sleep(Duration::from_millis(50)));
{
    let p = a.load(SeqCst, &guard);
    assert_eq!(unsafe { p.as_ref() }, Some(&777));
}

Trait Implementations

impl Drop for Guard

fn drop(&mut self)

Executes the destructor for this type.

impl Clone for Guard

fn clone(&self) -> Guard

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.