acca9e3834
The reasons for doing this are: * The model on which linked failure is based is inherently complex * The implementation is also very complex, and there are few remaining who fully understand the implementation * There are existing race conditions in the core context switching function of the scheduler, and possibly others. * It's unclear whether this model of linked failure maps well to a 1:1 threading model Linked failure is often a desired aspect of tasks, but we would like to take a much more conservative approach in re-implementing linked failure if at all. Closes #8674 Closes #8318 Closes #8863
93 lines
3.1 KiB
Rust
93 lines
3.1 KiB
Rust
// xfail-pretty
|
|
// xfail-test linked failure
|
|
|
|
// Copyright 2012 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.
|
|
|
|
/**
|
|
* Test performance of killing many tasks in a taskgroup.
|
|
* Along the way, tests various edge cases of ancestor group management.
|
|
* In particular, this tries to get each grandchild task to hit the
|
|
* "nobe_is_dead" case in each_ancestor only during task exit, but not during
|
|
* task spawn. This makes sure that defunct ancestor groups are handled correctly
|
|
* w.r.t. possibly leaving stale *rust_tasks lying around.
|
|
*/
|
|
|
|
// Creates in the background 'num_tasks' tasks, all blocked forever.
|
|
// Doesn't return until all such tasks are ready, but doesn't block forever itself.
|
|
|
|
use std::comm::{stream, SharedChan};
|
|
use std::os;
|
|
use std::result;
|
|
use std::task;
|
|
use std::uint;
|
|
|
|
fn grandchild_group(num_tasks: uint) {
|
|
let (po, ch) = stream();
|
|
let ch = SharedChan::new(ch);
|
|
|
|
for _ in range(0, num_tasks) {
|
|
let ch = ch.clone();
|
|
let mut t = task::task();
|
|
do t.spawn { // linked
|
|
ch.send(());
|
|
let (p, _c) = stream::<()>();
|
|
p.recv(); // block forever
|
|
}
|
|
}
|
|
error!("Grandchild group getting started");
|
|
for _ in range(0, num_tasks) {
|
|
// Make sure all above children are fully spawned; i.e., enlisted in
|
|
// their ancestor groups.
|
|
po.recv();
|
|
}
|
|
error!("Grandchild group ready to go.");
|
|
// Master grandchild task exits early.
|
|
}
|
|
|
|
fn spawn_supervised_blocking(myname: &str, f: proc()) {
|
|
let mut builder = task::task();
|
|
let res = builder.future_result();
|
|
builder.supervised();
|
|
builder.spawn(f);
|
|
error!("{} group waiting", myname);
|
|
let x = res.recv();
|
|
assert!(x.is_ok());
|
|
}
|
|
|
|
fn main() {
|
|
let args = os::args();
|
|
let args = if os::getenv("RUST_BENCH").is_some() {
|
|
~[~"", ~"100000"]
|
|
} else if args.len() <= 1u {
|
|
~[~"", ~"100"]
|
|
} else {
|
|
args.clone()
|
|
};
|
|
|
|
let num_tasks = from_str::<uint>(args[1]).unwrap();
|
|
|
|
// Main group #0 waits for unsupervised group #1.
|
|
// Grandparent group #1 waits for middle group #2, then fails, killing #3.
|
|
// Middle group #2 creates grandchild_group #3, waits for it to be ready, exits.
|
|
let x: result::Result<(), ~Any> = do task::try { // unlinked
|
|
do spawn_supervised_blocking("grandparent") {
|
|
do spawn_supervised_blocking("middle") {
|
|
grandchild_group(num_tasks);
|
|
}
|
|
// When grandchild group is ready to go, make the middle group exit.
|
|
error!("Middle group wakes up and exits");
|
|
}
|
|
// Grandparent group waits for middle group to be gone, then fails
|
|
error!("Grandparent group wakes up and fails");
|
|
fail!();
|
|
};
|
|
assert!(x.is_err());
|
|
}
|