task: update doc on spawn_blocking and block_in_place (#2436)

1 files changed, 62 insertions, 11 deletions

diff --git a/tokio/src/task/blocking.rs b/tokio/src/task/blocking.rs
index f9be1ead..0ef60535 100644
--- a/tokio/src/task/blocking.rs
+++ b/tokio/src/task/blocking.rs
@@ -1,20 +1,37 @@
 use crate::task::JoinHandle;
 
 cfg_rt_threaded! {
-    /// Runs the provided blocking function without blocking the executor.
+    /// Runs the provided blocking function on the current thread without
+    /// blocking the executor.
     ///
     /// In general, issuing a blocking call or performing a lot of compute in a
     /// future without yielding is not okay, as it may prevent the executor from
-    /// driving other futures forward. If you run a closure through this method,
-    /// the current executor thread will relegate all its executor duties to another
-    /// (possibly new) thread, and only then poll the task. Note that this requires
-    /// additional synchronization.
+    /// driving other futures forward.  This function runs the closure on the
+    /// current thread by having the thread temporarily cease from being a core
+    /// thread, and turns it into a blocking thread. See the [CPU-bound tasks
+    /// and blocking code][blocking] section for more information.
     ///
-    /// # Note
+    /// Although this function avoids starving other independently spawned
+    /// tasks, any other code running concurrently in the same task will be
+    /// suspended during the call to `block_in_place`. This can happen e.g. when
+    /// using the [`join!`] macro. To avoid this issue, use [`spawn_blocking`]
+    /// instead.
     ///
-    /// This function can only be called from a spawned task when working with
-    /// the [threaded scheduler](https://docs.rs/tokio/0.2.10/tokio/runtime/index.html#threaded-scheduler).
-    /// Consider using [tokio::task::spawn_blocking](https://docs.rs/tokio/0.2.10/tokio/task/fn.spawn_blocking.html).
+    /// Note that this function can only be used on the [threaded scheduler].
+    ///
+    /// Code running behind `block_in_place` cannot be cancelled. When you shut
+    /// down the executor, it will wait indefinitely for all blocking operations
+    /// to finish. You can use [`shutdown_timeout`] to stop waiting for them
+    /// after a certain timeout. Be aware that this will still not cancel the
+    /// tasks — they are simply allowed to keep running after the method
+    /// returns.
+    ///
+    /// [blocking]: ../index.html#cpu-bound-tasks-and-blocking-code
+    /// [threaded scheduler]: fn@crate::runtime::Builder::threaded_scheduler
+    /// [`spawn_blocking`]: fn@crate::task::spawn_blocking
+    /// [`join!`]: ../macro.join.html
+    /// [`thread::spawn`]: fn@std::thread::spawn
+    /// [`shutdown_timeout`]: fn@crate::runtime::Runtime::shutdown_timeout
     ///
     /// # Examples
     ///
@@ -39,9 +56,43 @@ cfg_rt_threaded! {
 cfg_blocking! {
     /// Runs the provided closure on a thread where blocking is acceptable.
     ///
-    /// More information about the blocking threads is available at [CPU-bound tasks and blocking code][blocking].
+    /// In general, issuing a blocking call or performing a lot of compute in a
+    /// future without yielding is not okay, as it may prevent the executor from
+    /// driving other futures forward. This function runs the provided closure
+    /// on a thread dedicated to blocking operations. See the [CPU-bound tasks
+    /// and blocking code][blocking] section for more information.
+    ///
+    /// Tokio will spawn more blocking threads when they are requested through
+    /// this function until the upper limit configured on the [`Builder`] is
+    /// reached.  This limit is very large by default, because `spawn_blocking` is
+    /// often used for various kinds of IO operations that cannot be performed
+    /// asynchronously. When you run CPU-bound code using `spawn_blocking`, you
+    /// should keep this large upper limit in mind; to run your CPU-bound
+    /// computations on only a few threads, you should use a separate thread
+    /// pool such as [rayon] rather than configuring the number of blocking
+    /// threads.
+    ///
+    /// This function is intended for non-async operations that eventually
+    /// finish on their own. If you want to spawn an ordinary thread, you should
+    /// use [`thread::spawn`] instead.
+    ///
+    /// Closures spawned using `spawn_blocking` cannot be cancelled. When you
+    /// shut down the executor, it will wait indefinitely for all blocking
+    /// operations to finish. You can use [`shutdown_timeout`] to stop waiting
+    /// for them after a certain timeout. Be aware that this will still not
+    /// cancel the tasks — they are simply allowed to keep running after the
+    /// method returns.
+    ///
+    /// Note that if you are using the [basic scheduler], this function will
+    /// still spawn additional threads for blocking operations. The basic
+    /// scheduler's single thread is only used for asynchronous code.
     ///
-    ///  [blocking]: ../index.html#cpu-bound-tasks-and-blocking-code
+    /// [`Builder`]: struct@crate::runtime::Builder
+    /// [blocking]: ../index.html#cpu-bound-tasks-and-blocking-code
+    /// [rayon]: https://docs.rs/rayon
+    /// [basic scheduler]: fn@crate::runtime::Builder::basic_scheduler
+    /// [`thread::spawn`]: fn@std::thread::spawn
+    /// [`shutdown_timeout`]: fn@crate::runtime::Runtime::shutdown_timeout
     ///
     /// # Examples
     ///