auto merge of #15944 : alexcrichton/rust/task-dont-die, r=brson

Previously both spawning mechanisms were not resilient to task failures which were initiated from the task spawning infrastructure.

Closes #15895

Author: bors

src/libgreen/sched.rs

@@ -219,7 +219,7 @@ impl Scheduler {
         let message = stask.sched.get_mut_ref().message_queue.pop();
         rtassert!(match message { msgq::Empty => true, _ => false });
 
-        stask.task.get_mut_ref().destroyed = true;
+        stask.task.take().unwrap().drop();
     }
 
     // This does not return a scheduler, as the scheduler is placed

src/libgreen/task.rs

@@ -442,15 +442,30 @@ impl Runtime for GreenTask {
                      f: proc():Send) {
         self.put_task(cur_task);
 
+        // First, set up a bomb which when it goes off will restore the local
+        // task unless its disarmed. This will allow us to gracefully fail from
+        // inside of `configure` which allocates a new task.
+        struct Bomb { inner: Option<Box<GreenTask>> }
+        impl Drop for Bomb {
+            fn drop(&mut self) {
+                let _ = self.inner.take().map(|task| task.put());
+            }
+        }
+        let mut bomb = Bomb { inner: Some(self) };
+
         // Spawns a task into the current scheduler. We allocate the new task's
         // stack from the scheduler's stack pool, and then configure it
         // accordingly to `opts`. Afterwards we bootstrap it immediately by
         // switching to it.
         //
         // Upon returning, our task is back in TLS and we're good to return.
-        let mut sched = self.sched.take_unwrap();
-        let sibling = GreenTask::configure(&mut sched.stack_pool, opts, f);
-        sched.run_task(self, sibling)
+        let sibling = {
+            let sched = bomb.inner.get_mut_ref().sched.get_mut_ref();
+            GreenTask::configure(&mut sched.stack_pool, opts, f)
+        };
+        let mut me = bomb.inner.take().unwrap();
+        let sched = me.sched.take().unwrap();
+        sched.run_task(me, sibling)
     }
 
     // Local I/O is provided by the scheduler's event loop

src/libnative/task.rs

@@ -71,7 +71,7 @@ pub fn spawn_opts(opts: TaskOpts, f: proc():Send) {
     // Note that this increment must happen *before* the spawn in order to
     // guarantee that if this task exits it will always end up waiting for the
     // spawned task to exit.
-    bookkeeping::increment();
+    let token = bookkeeping::increment();
 
     // Spawning a new OS thread guarantees that __morestack will never get
     // triggered, but we must manually set up the actual stack bounds once this
@@ -93,7 +93,7 @@ pub fn spawn_opts(opts: TaskOpts, f: proc():Send) {
         let mut task = task;
         task.put_runtime(ops);
         drop(task.run(|| { f.take_unwrap()() }).destroy());
-        bookkeeping::decrement();
+        drop(token);
     })
 }
 

src/librustrt/bookkeeping.rs

@@ -19,14 +19,24 @@
 //! decrement() manually.
 
 use core::atomics;
+use core::ops::Drop;
 
 use mutex::{StaticNativeMutex, NATIVE_MUTEX_INIT};
 
 static mut TASK_COUNT: atomics::AtomicUint = atomics::INIT_ATOMIC_UINT;
 static mut TASK_LOCK: StaticNativeMutex = NATIVE_MUTEX_INIT;
 
-pub fn increment() {
+pub struct Token { _private: () }
+
+impl Drop for Token {
+    fn drop(&mut self) { decrement() }
+}
+
+/// Increment the number of live tasks, returning a token which will decrement
+/// the count when dropped.
+pub fn increment() -> Token {
     let _ = unsafe { TASK_COUNT.fetch_add(1, atomics::SeqCst) };
+    Token { _private: () }
 }
 
 pub fn decrement() {

src/librustrt/local.rs

@@ -125,8 +125,8 @@ mod test {
         }).join();
     }
 
-    fn cleanup_task(mut t: Box<Task>) {
-        t.destroyed = true;
+    fn cleanup_task(t: Box<Task>) {
+        t.drop();
     }
 
 }

src/librustrt/task.rs

@@ -100,12 +100,21 @@ pub struct Task {
     pub storage: LocalStorage,
     pub unwinder: Unwinder,
     pub death: Death,
-    pub destroyed: bool,
     pub name: Option<SendStr>,
 
+    state: TaskState,
     imp: Option<Box<Runtime + Send>>,
 }
 
+// Once a task has entered the `Armed` state it must be destroyed via `drop`,
+// and no other method. This state is used to track this transition.
+#[deriving(PartialEq)]
+enum TaskState {
+    New,
+    Armed,
+    Destroyed,
+}
+
 pub struct TaskOpts {
     /// Invoke this procedure with the result of the task when it finishes.
     pub on_exit: Option<proc(Result): Send>,
@@ -159,7 +168,7 @@ impl Task {
             storage: LocalStorage(None),
             unwinder: Unwinder::new(),
             death: Death::new(),
-            destroyed: false,
+            state: New,
             name: None,
             imp: None,
         }
@@ -203,7 +212,7 @@ impl Task {
     /// }).destroy();
     /// # }
     /// ```
-    pub fn run(self: Box<Task>, f: ||) -> Box<Task> {
+    pub fn run(mut self: Box<Task>, f: ||) -> Box<Task> {
         assert!(!self.is_destroyed(), "cannot re-use a destroyed task");
 
         // First, make sure that no one else is in TLS. This does not allow
@@ -212,6 +221,7 @@ impl Task {
         if Local::exists(None::<Task>) {
             fail!("cannot run a task recursively inside another");
         }
+        self.state = Armed;
         Local::put(self);
 
         // There are two primary reasons that general try/catch is unsafe. The
@@ -333,12 +343,12 @@ impl Task {
         // Now that we're done, we remove the task from TLS and flag it for
         // destruction.
         let mut task: Box<Task> = Local::take();
-        task.destroyed = true;
+        task.state = Destroyed;
         return task;
     }
 
     /// Queries whether this can be destroyed or not.
-    pub fn is_destroyed(&self) -> bool { self.destroyed }
+    pub fn is_destroyed(&self) -> bool { self.state == Destroyed }
 
     /// Inserts a runtime object into this task, transferring ownership to the
     /// task. It is illegal to replace a previous runtime object in this task
@@ -453,12 +463,20 @@ impl Task {
     pub fn can_block(&self) -> bool {
         self.imp.get_ref().can_block()
     }
+
+    /// Consume this task, flagging it as a candidate for destruction.
+    ///
+    /// This function is required to be invoked to destroy a task. A task
+    /// destroyed through a normal drop will abort.
+    pub fn drop(mut self) {
+        self.state = Destroyed;
+    }
 }
 
 impl Drop for Task {
     fn drop(&mut self) {
         rtdebug!("called drop for a task: {}", self as *mut Task as uint);
-        rtassert!(self.destroyed);
+        rtassert!(self.state != Armed);
     }
 }
 
@@ -634,12 +652,17 @@ mod test {
         begin_unwind("cause", file!(), line!())
     }
 
+    #[test]
+    fn drop_new_task_ok() {
+        drop(Task::new());
+    }
+
     // Task blocking tests
 
     #[test]
     fn block_and_wake() {
         let task = box Task::new();
         let mut task = BlockedTask::block(task).wake().unwrap();
-        task.destroyed = true;
+        task.drop();
     }
 }

src/test/run-pass/spawn-stack-too-big.rs

@@ -0,0 +1,47 @@
+// Copyright 2014 The Rust Project Developers. See the COPYRIGHT
+// file at the top-level directory of this distribution and at
+// http://rust-lang.org/COPYRIGHT.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+// ignore-macos apparently gargantuan mmap requests are ok?
+
+#![feature(phase)]
+
+#[phase(plugin)]
+extern crate green;
+extern crate native;
+
+use std::task::TaskBuilder;
+use native::NativeTaskBuilder;
+
+green_start!(main)
+
+fn main() {
+    test();
+
+    let (tx, rx) = channel();
+    TaskBuilder::new().native().spawn(proc() {
+        tx.send(test());
+    });
+    rx.recv();
+}
+
+#[cfg(not(target_word_size = "64"))]
+fn test() {}
+
+#[cfg(target_word_size = "64")]
+fn test() {
+    let (tx, rx) = channel();
+    spawn(proc() {
+        TaskBuilder::new().stack_size(1024 * 1024 * 1024 * 64).spawn(proc() {
+        });
+        tx.send(());
+    });
+
+    assert!(rx.recv_opt().is_err());
+}