add nonpoison::once implementation

Adds the equivalent `nonpoison` types to the `poison::once` module.
These types and implementations are gated under the `nonpoison_once`
feature gate.

Author: connortsui20

library/std/src/sync/nonpoison.rs

@@ -33,5 +33,8 @@ impl fmt::Display for WouldBlock {
 pub use self::mutex::MappedMutexGuard;
 #[unstable(feature = "nonpoison_mutex", issue = "134645")]
 pub use self::mutex::{Mutex, MutexGuard};
+#[unstable(feature = "nonpoison_once", issue = "134645")]
+pub use self::once::Once;
 
 mod mutex;
+pub(crate) mod once;

library/std/src/sync/nonpoison/once.rs

@@ -0,0 +1,221 @@
+use crate::fmt;
+use crate::sys::sync as sys;
+
+/// A low-level synchronization primitive for one-time global execution that ignores poisoning.
+///
+/// For more information about `Once`, check out the documentation for the poisoning variant (which
+/// can be found at [`poison::Once`]).
+///
+/// [`poison::Once`]: crate::sync::poison::Once
+///
+/// # Examples
+///
+/// ```
+/// use std::sync::Once;
+///
+/// static START: Once = Once::new();
+///
+/// START.call_once(|| {
+///     // run initialization here
+/// });
+/// ```
+#[unstable(feature = "nonpoison_once", issue = "134645")]
+pub struct Once {
+    inner: sys::Once,
+}
+
+impl Once {
+    /// Creates a new `Once` value.
+    #[inline]
+    #[unstable(feature = "nonpoison_once", issue = "134645")]
+    #[must_use]
+    pub const fn new() -> Once {
+        Once { inner: sys::Once::new() }
+    }
+
+    /// Performs an initialization routine once and only once. The given closure
+    /// will be executed if this is the first time `call_once` has been called,
+    /// and otherwise the routine will *not* be invoked.
+    ///
+    /// This method will block the calling thread if another initialization
+    /// routine is currently running.
+    ///
+    /// When this function returns, it is guaranteed that some initialization
+    /// has run and completed (it might not be the closure specified). It is also
+    /// guaranteed that any memory writes performed by the executed closure can
+    /// be reliably observed by other threads at this point (there is a
+    /// happens-before relation between the closure and code executing after the
+    /// return).
+    ///
+    /// If the given closure recursively invokes `call_once` on the same [`Once`]
+    /// instance, the exact behavior is not specified: allowed outcomes are
+    /// a panic or a deadlock.
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// #![feature(nonpoison_once)]
+    ///
+    /// use std::sync::nonpoison::Once;
+    ///
+    /// static mut VAL: usize = 0;
+    /// static INIT: Once = Once::new();
+    ///
+    /// // Accessing a `static mut` is unsafe much of the time, but if we do so
+    /// // in a synchronized fashion (e.g., write once or read all) then we're
+    /// // good to go!
+    /// //
+    /// // This function will only call `expensive_computation` once, and will
+    /// // otherwise always return the value returned from the first invocation.
+    /// fn get_cached_val() -> usize {
+    ///     unsafe {
+    ///         INIT.call_once(|| {
+    ///             VAL = expensive_computation();
+    ///         });
+    ///         VAL
+    ///     }
+    /// }
+    ///
+    /// fn expensive_computation() -> usize {
+    ///     // ...
+    /// # 2
+    /// }
+    /// ```
+    ///
+    /// # Panics
+    ///
+    /// The closure `f` will only be executed once even if this is called concurrently amongst many
+    /// threads. If the closure panics, the calling thread will panic and the `Once` will remain in
+    /// an incompleted state.
+    ///
+    /// In contrast to the [`poison::Once`] variant, all calls to `call_once` will ignore panics in
+    /// other threads. This method is identical to the [`poison::Once::call_once_force`] method.
+    ///
+    /// If you need observability into whether any threads have panicked while calling `call_once`,
+    /// see [`poison::Once`].
+    ///
+    /// [`poison::Once`]: crate::sync::poison::Once
+    /// [`poison::Once::call_once_force`]: crate::sync::poison::Once::call_once_force
+    ///
+    /// ```
+    /// #![feature(nonpoison_once)]
+    ///
+    /// use std::sync::nonpoison::Once;
+    /// use std::thread;
+    ///
+    /// static INIT: Once = Once::new();
+    ///
+    /// // Panic during `call_once`.
+    /// let handle = thread::spawn(|| {
+    ///     INIT.call_once(|| panic!());
+    /// });
+    /// assert!(handle.join().is_err());
+    ///
+    /// // `call_once` will still run from a different thread.
+    /// INIT.call_once(|| {
+    ///     assert_eq!(2 + 2, 4);
+    /// });
+    /// ```
+    #[inline]
+    #[unstable(feature = "nonpoison_once", issue = "134645")]
+    #[track_caller]
+    pub fn call_once<F>(&self, f: F)
+    where
+        F: FnOnce(),
+    {
+        // Fast path check.
+        if self.inner.is_completed() {
+            return;
+        }
+
+        let mut f = Some(f);
+        self.inner.call(true, &mut |_| f.take().unwrap()());
+    }
+
+    /// Returns `true` if some [`call_once()`] call has completed
+    /// successfully. Specifically, `is_completed` will return false in
+    /// the following situations:
+    ///   * [`call_once()`] was not called at all,
+    ///   * [`call_once()`] was called, but has not yet completed
+    ///
+    /// This function returning `false` does not mean that [`Once`] has not been
+    /// executed. For example, it may have been executed in the time between
+    /// when `is_completed` starts executing and when it returns, in which case
+    /// the `false` return value would be stale (but still permissible).
+    ///
+    /// [`call_once()`]: Once::call_once
+    ///
+    /// # Examples
+    ///
+    /// ```
+    /// #![feature(nonpoison_once)]
+    ///
+    /// use std::sync::nonpoison::Once;
+    ///
+    /// static INIT: Once = Once::new();
+    ///
+    /// assert_eq!(INIT.is_completed(), false);
+    /// INIT.call_once(|| {
+    ///     assert_eq!(INIT.is_completed(), false);
+    /// });
+    /// assert_eq!(INIT.is_completed(), true);
+    /// ```
+    ///
+    /// ```
+    /// #![feature(nonpoison_once)]
+    ///
+    /// use std::sync::nonpoison::Once;
+    /// use std::thread;
+    ///
+    /// static INIT: Once = Once::new();
+    ///
+    /// assert_eq!(INIT.is_completed(), false);
+    /// let handle = thread::spawn(|| {
+    ///     INIT.call_once(|| panic!());
+    /// });
+    /// assert!(handle.join().is_err());
+    /// assert_eq!(INIT.is_completed(), false);
+    /// ```
+    #[inline]
+    #[unstable(feature = "nonpoison_once", issue = "134645")]
+    pub fn is_completed(&self) -> bool {
+        self.inner.is_completed()
+    }
+
+    /// Blocks the current thread until initialization has completed.
+    ///
+    /// # Example
+    ///
+    /// ```rust
+    /// use std::sync::Once;
+    /// use std::thread;
+    ///
+    /// static READY: Once = Once::new();
+    ///
+    /// let thread = thread::spawn(|| {
+    ///     READY.wait();
+    ///     println!("everything is ready");
+    /// });
+    ///
+    /// READY.call_once(|| println!("performing setup"));
+    /// ```
+    ///
+    /// This function will continue to block even if a thread initializing via [`call_once()`] has
+    /// panicked. This behavior is identical to the [`poison::Once::wait_force`] method.
+    ///
+    /// [`call_once()`]: Once::call_once
+    /// [`poison::Once::wait_force`]: crate::sync::poison::Once::wait_force
+    #[unstable(feature = "nonpoison_once", issue = "134645")]
+    pub fn wait(&self) {
+        if !self.inner.is_completed() {
+            self.inner.wait(true);
+        }
+    }
+}
+
+#[unstable(feature = "nonpoison_once", issue = "134645")]
+impl fmt::Debug for Once {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.debug_struct("Once").finish_non_exhaustive()
+    }
+}