Struct cros_async::CancellableBlockingPool

source ·

pub struct CancellableBlockingPool {
    inner: Arc<Inner>,
}

Expand description

A thread pool for running work that may block.

This is a wrapper around BlockingPool with an ability to cancel queued tasks. See BlockingPool for more info.

§Examples

Spawn a task to run in the CancellableBlockingPool and await on its result.

use cros_async::CancellableBlockingPool;

    let pool = CancellableBlockingPool::default();
    let CANCELLED = 0;

    let res = pool.spawn(move || {
        // Do some CPU-intensive or blocking work here.

        42
    }, move || CANCELLED).await;

    assert_eq!(res, 42);

Fields§

§inner: Arc<Inner>

Implementations§

source §

pub fn new(max_threads: usize, keepalive: Duration) -> CancellableBlockingPool

Create a new CancellableBlockingPool.

When we try to shutdown or drop CancellableBlockingPool, it may happen that a hung thread might prevent CancellableBlockingPool pool from getting dropped. On failure to shutdown in watchdog_opts.timeout duration, CancellableBlockingPool can take an action specified by watchdog_opts.action.

pub fn with_capacity( max_threads: usize, keepalive: Duration, ) -> CancellableBlockingPool

Like Self::new but with pre-allocating capacity for up to max_threads.

source

pub fn spawn<F, R, G>(&self, f: F, cancel: G) -> impl Future<Output = R>
where F: FnOnce() -> R + Send + 'static, R: Send + 'static, G: Fn() -> R + Send + 'static,

Spawn a task to run in the CancellableBlockingPool.

Callers may await the returned Task to be notified when the work is completed. Dropping the future will not cancel the task.

cancel helps to cancel a queued or in-flight operation f. cancel may be called more than once if f doesn’t respond to cancel. cancel is not called if f completes successfully. For example,

§Examples

use {cros_async::CancellableBlockingPool, std::sync::{Arc, Mutex, Condvar}};

   let pool = CancellableBlockingPool::default();
   let cancelled: i32 = 1;
   let success: i32 = 2;

   let shared = Arc::new((Mutex::new(0), Condvar::new()));
   let shared2 = shared.clone();
   let shared3 = shared.clone();

   let res = pool
       .spawn(
           move || {
               let guard = shared.0.lock().unwrap();
               let mut guard = shared.1.wait_while(guard, |state| *state == 0).unwrap();
               if *guard != cancelled {
                   *guard = success;
               }
           },
           move || {
               *shared2.0.lock().unwrap() = cancelled;
               shared2.1.notify_all();
           },
       )
       .await;
   pool.shutdown();

   assert_eq!(*shared3.0.lock().unwrap(), cancelled);

source

fn drain_cancellables(&self)

Iterates over all the queued tasks and marks them as cancelled.

source

pub fn disarm(&self)

Marks all the queued and in-flight tasks as cancelled. Any tasks queued after disarming will be cancelled. Does not wait for all the tasks to get cancelled.

source

pub fn shutdown(&self) -> Result<(), Error>

Shut down the CancellableBlockingPool.

This will block until all work that has been started by the worker threads is finished. Any work that was added to the CancellableBlockingPool but not yet picked up by a worker thread will not complete and awaiting on the Task for that work will panic.

source