feat(link): bidirectional links + trap_exit (roadmap #3)

Add Erlang-style process links so an abnormal death fate-shares across a
link set, with trap_exit to convert that into a message instead.

- Slot.links: Vec<Pid>, bidirectional; reset in all three lifecycle sites.
- Actor.trap: Option<Sender<ExitSignal>>, fresh per spawn (a restarted
  child starts un-trapped; no fourth reset site).
- link/unlink free functions on self_pid(); trap_exit() -> Receiver<ExitSignal>
  (a dedicated inbox, distinct from the monitor Down channel).
- finalize_actor clears reverse links under the lock (always; keeps the
  cascade acyclic), then propagates abnormal deaths post-lock: trapping peer
  gets an ExitSignal message and survives, non-trapping peer is request_stop'd.
  Normal exit never propagates. Dead-pid link delivers an immediate NoProc
  (message if trapping, else request_stop(self) -- never a silent no-op).
- ExitSignal reuses DownReason and carries no panic payload (joiner-only).

Tests in tests/link.rs cover the propagate/trap/normal/dead-pid/unlink cases.
This commit is contained in:
smarm-dev
2026-06-07 21:51:56 +00:00
parent 59998ce79e
commit 8ac11a57ac
6 changed files with 443 additions and 2 deletions
+7
View File
@@ -134,4 +134,11 @@ pub struct Actor {
/// (constructed fresh at spawn), so unlike a `Slot` field it needs no reset
/// in the slot-recycling paths.
pub stop: std::sync::Arc<std::sync::atomic::AtomicBool>,
/// Trap-exit inbox (roadmap #3). `None` until the actor calls
/// `link::trap_exit()`; `Some(tx)` means an abnormal death of a linked
/// peer is delivered here as an `ExitSignal` *message* instead of
/// cooperatively stopping this actor. Lives on the `Actor` (fresh per
/// spawn) for the same reason as `stop`: no slot-recycling reset needed,
/// and a restarted child correctly starts out *not* trapping.
pub trap: Option<crate::channel::Sender<crate::link::ExitSignal>>,
}
+2
View File
@@ -23,6 +23,7 @@ pub mod timer;
pub mod io;
pub mod mutex;
pub mod monitor;
pub mod link;
pub mod runtime;
pub mod trace;
@@ -38,6 +39,7 @@ static ALLOCATOR: preempt::PreemptingAllocator = preempt::PreemptingAllocator;
// ---------------------------------------------------------------------------
pub use channel::{channel, Receiver, RecvError, Sender};
pub use link::{link, trap_exit, unlink, ExitSignal};
pub use monitor::{monitor, Down, DownReason};
pub use mutex::{LockTimeout, Mutex, MutexGuard};
pub use pid::Pid;
+164
View File
@@ -0,0 +1,164 @@
//! Process links + `trap_exit`.
//!
//! A *link* is bidirectional and persistent, in contrast to a [`monitor`]
//! (unidirectional, one-shot): once two actors are linked, the *abnormal*
//! death of either propagates to the other. "Abnormal" means a panic or a
//! cooperative [`request_stop`] — a normal return never propagates.
//!
//! What propagation *does* depends on whether the surviving peer traps exits:
//!
//! - **not trapping** (the default): the peer is cooperatively
//! [`request_stop`]'d, so a crash fate-shares across the whole link set —
//! Erlang's "let it crash". Because the peer's own links are walked when
//! *it* finalizes, the stop cascades transitively.
//! - **trapping** (called [`trap_exit`]): the peer instead receives an
//! [`ExitSignal`] *message* on its trap inbox and keeps running, turning
//! peer failures into ordinary inbox events (Erlang's
//! `process_flag(trap_exit, true)`).
//!
//! Linking an already-dead pid is not a silent no-op: it behaves exactly like
//! an immediate abnormal death of the target with reason [`DownReason::NoProc`]
//! — a trapping caller gets a message, a non-trapping caller is stopped.
//!
//! ## Payload
//!
//! [`ExitSignal`] reuses [`DownReason`] and, like a monitor, does **not**
//! carry the panic payload: a panic's payload has a single owner and is
//! delivered to whoever `join()`s the actor. A trap inbox only learns *that*
//! (and *how*) a peer died.
//!
//! ## Inbox
//!
//! The trap inbox is a dedicated channel, distinct from the monitor [`Down`]
//! channel: monitors are documented as one-shot/unidirectional, whereas a trap
//! inbox receives one [`ExitSignal`] per linked peer death over the actor's
//! lifetime. [`trap_exit`] enables trapping and hands back the inbox in one
//! call; calling it again installs a fresh inbox (the previous one closes).
//!
//! ## Races
//!
//! [`link`] registration and `finalize_actor` (in `runtime`) both serialize on
//! the shared mutex, so a target that is live when the link is registered is
//! guaranteed to propagate its eventual death; there is no window in which the
//! death slips between the liveness check and the registration. As elsewhere,
//! we never `send` or `request_stop` while holding the shared lock — those
//! re-enter the runtime — so the act is always performed after the lock is
//! released.
//!
//! [`monitor`]: crate::monitor::monitor
//! [`Down`]: crate::monitor::Down
//! [`request_stop`]: crate::scheduler::request_stop
use crate::channel::{channel, Receiver, Sender};
use crate::monitor::DownReason;
use crate::pid::Pid;
use crate::runtime::State;
use crate::scheduler::{request_stop, self_pid, with_runtime};
/// A linked peer's death, delivered to a trapping actor's inbox.
///
/// `Copy` because it carries no payload — see the module docs for why the
/// panic payload is *not* included.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub struct ExitSignal {
/// The linked pid that died (or the dead pid that was linked).
pub from: Pid,
/// How it went down. Only ever `Panic`, `Stopped`, or `NoProc` — a normal
/// `Exit` does not propagate, so it never appears here.
pub reason: DownReason,
}
/// Trap exits on the current actor and return its exit inbox.
///
/// After this call, the abnormal death of a linked peer arrives here as an
/// [`ExitSignal`] message instead of cooperatively stopping this actor.
/// Calling it again installs a fresh inbox; the previously returned receiver
/// then closes.
pub fn trap_exit() -> Receiver<ExitSignal> {
let (tx, rx) = channel::<ExitSignal>();
let me = self_pid();
with_runtime(|inner| {
inner.with_shared(|s| {
if let Some(actor) = s.slot_mut(me).and_then(|slot| slot.actor.as_mut()) {
actor.trap = Some(tx);
}
})
});
rx
}
/// Bidirectionally link the current actor to `target`.
///
/// If `target` is live, the link is recorded on both actors and either's
/// abnormal death will propagate to the other. If `target` is already gone,
/// this delivers an immediate [`DownReason::NoProc`] exit signal to the caller
/// (a message if trapping, otherwise a cooperative stop). Linking yourself, or
/// re-linking an existing peer, is a no-op.
pub fn link(target: Pid) {
let me = self_pid();
if target == me {
return;
}
// Under the lock: if the target is live, record the link both ways and
// return `None`. If it is gone, return the caller's trap sender (if any)
// so we can deliver the NoProc signal after releasing the lock.
let dead_action: Option<Option<Sender<ExitSignal>>> = with_runtime(|inner| {
inner.with_shared(|s| {
let target_live = matches!(
s.slot(target),
Some(slot) if slot.actor.is_some() && !matches!(slot.state, State::Done)
);
if target_live {
if let Some(slot) = s.slot_mut(me) {
if !slot.links.contains(&target) {
slot.links.push(target);
}
}
if let Some(slot) = s.slot_mut(target) {
if !slot.links.contains(&me) {
slot.links.push(me);
}
}
None
} else {
// Grab our own trap sender so the NoProc delivery (below)
// doesn't need a second lock acquisition.
Some(
s.slot(me)
.and_then(|slot| slot.actor.as_ref())
.and_then(|a| a.trap.clone()),
)
}
})
});
match dead_action {
None => {} // linked successfully
Some(Some(tx)) => {
let _ = tx.send(ExitSignal { from: target, reason: DownReason::NoProc });
}
Some(None) => request_stop(me),
}
}
/// Remove the link between the current actor and `target`, in both directions.
///
/// After this, neither actor's death propagates to the other. A no-op if the
/// two were not linked.
pub fn unlink(target: Pid) {
let me = self_pid();
if target == me {
return;
}
with_runtime(|inner| {
inner.with_shared(|s| {
if let Some(slot) = s.slot_mut(me) {
slot.links.retain(|p| *p != target);
}
if let Some(slot) = s.slot_mut(target) {
slot.links.retain(|p| *p != me);
}
})
});
}
+45 -2
View File
@@ -228,6 +228,11 @@ pub(crate) struct Slot {
/// actor terminates (drained in `finalize_actor`). Distinct from
/// `supervisor_channel`, which is the parent's single funnel.
pub(crate) monitors: Vec<Sender<Down>>,
/// Bidirectional links (roadmap #3). Each entry is a peer whose abnormal
/// death propagates to this actor (and vice versa — the relationship is
/// recorded on both slots). Walked + cleared in `finalize_actor`. Reset in
/// all three slot-lifecycle sites, like `monitors`.
pub(crate) links: Vec<Pid>,
pub(crate) outstanding_handles: u32,
pub(crate) pending_io_result: Option<crate::io::IoResult>,
/// Set by `unpark()` when the actor is still running (not yet Parked).
@@ -246,6 +251,7 @@ impl Slot {
outcome: None,
supervisor_channel: None,
monitors: Vec::new(),
links: Vec::new(),
outstanding_handles: 0,
pending_io_result: None,
pending_unpark: false,
@@ -523,6 +529,7 @@ pub(crate) fn reclaim_slot(s: &mut SharedState, pid: Pid) {
slot.waiters.clear();
slot.supervisor_channel = None;
slot.monitors.clear();
slot.links.clear();
slot.state = State::Done;
slot.outstanding_handles = 0;
slot.pending_unpark = false;
@@ -548,7 +555,7 @@ fn finalize_actor(inner: &Arc<RuntimeInner>, pid: Pid, outcome: Outcome) {
Outcome::Stopped => (Outcome::Stopped, Signal::Stopped(pid), DownReason::Stopped),
};
let (waiters, supervisor_pid, monitors, recycled_stack) = inner.with_shared(|s| {
let (waiters, supervisor_pid, monitors, recycled_stack, links) = inner.with_shared(|s| {
let slot = s.slot_mut(pid).expect("finalize_actor: slot vanished");
let sup = slot.actor.as_ref().map(|a| a.supervisor);
// Extract the stack before clearing the actor so we can recycle it
@@ -557,7 +564,18 @@ fn finalize_actor(inner: &Arc<RuntimeInner>, pid: Pid, outcome: Outcome) {
slot.outcome = Some(joiner_outcome);
slot.state = State::Done;
let monitors = std::mem::take(&mut slot.monitors);
(std::mem::take(&mut slot.waiters), sup, monitors, stack)
let waiters = std::mem::take(&mut slot.waiters);
let links = std::mem::take(&mut slot.links);
// Drop this (dying) pid from each linked peer's list now, under the
// same lock. Done regardless of how we died, so a peer that later
// finalizes won't try to propagate back to this about-to-be-reclaimed
// slot — which also makes the abnormal-death cascade acyclic.
for &peer in &links {
if let Some(ps) = s.slot_mut(peer) {
ps.links.retain(|p| *p != pid);
}
}
(waiters, sup, monitors, stack, links)
});
// Return the stack to the pool outside the shared lock. The pool grows
@@ -588,6 +606,31 @@ fn finalize_actor(inner: &Arc<RuntimeInner>, pid: Pid, outcome: Outcome) {
let _ = m.send(Down { pid, reason: down_reason });
}
// Propagate to linked peers. A normal exit never propagates; an abnormal
// death (panic or cooperative stop) reaches each linked peer — a trapping
// peer gets an `ExitSignal` message and survives, a non-trapping peer is
// cooperatively stopped (which cascades along *its* links in turn). The
// reverse links were cleared above, so this never ping-pongs back. Decide
// trap-vs-stop under the lock, then act after releasing it: both
// `Sender::send` and `request_stop` re-enter the shared mutex.
if matches!(down_reason, DownReason::Panic | DownReason::Stopped) {
for peer in links {
let trap = inner.with_shared(|s| match s.slot(peer) {
Some(slot) if !matches!(slot.state, State::Done) => {
Some(slot.actor.as_ref().and_then(|a| a.trap.clone()))
}
_ => None, // peer already gone; nothing to do
});
match trap {
Some(Some(tx)) => {
let _ = tx.send(crate::link::ExitSignal { from: pid, reason: down_reason });
}
Some(None) => crate::scheduler::request_stop(peer),
None => {}
}
}
}
// Unpark joiners.
for joiner in waiters {
crate::scheduler::unpark(joiner);
+2
View File
@@ -157,6 +157,7 @@ pub fn spawn_under(supervisor: Pid, f: impl FnOnce() + Send + 'static) -> JoinHa
sp,
supervisor,
stop: std::sync::Arc::new(std::sync::atomic::AtomicBool::new(false)),
trap: None,
});
slot.state = crate::runtime::State::Runnable;
slot.outstanding_handles = 1;
@@ -164,6 +165,7 @@ pub fn spawn_under(supervisor: Pid, f: impl FnOnce() + Send + 'static) -> JoinHa
slot.waiters.clear();
slot.supervisor_channel = None;
slot.monitors.clear();
slot.links.clear();
slot.pending_unpark = false;
slot.pending_io_result = None;
s.run_queue.push_back(pid);
+223
View File
@@ -0,0 +1,223 @@
//! Link + trap_exit tests (roadmap item #3).
//!
//! A *link* is bidirectional and persistent (contrast a monitor, which is
//! unidirectional and one-shot). When a linked actor dies *abnormally* — a
//! panic or a cooperative `request_stop` — the death propagates to the peer:
//!
//! - a peer that has NOT trapped exits is cooperatively `request_stop`'d, so
//! a crash fate-shares across the link set (Erlang's "let it crash");
//! - a peer that HAS called `trap_exit()` instead receives an [`ExitSignal`]
//! *message* on its trap inbox and survives.
//!
//! A *normal* exit never propagates — not even to a trapping peer. Linking an
//! already-dead pid delivers an immediate exit signal (`NoProc`): a message to
//! a trapping caller, a `request_stop` to a non-trapping one.
//!
//! These cases are deterministic under the single-thread runtime: an actor
//! runs until it parks/yields, so the relative order of link setup, the
//! triggering death, and the propagated signal is fixed.
use smarm::{
channel, link, monitor, run, self_pid, spawn, trap_exit, unlink, yield_now, DownReason,
};
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::Arc;
/// Sets its flag when dropped — proves a propagated stop unwound the peer's
/// stack (running Drop) rather than leaking it.
struct DropFlag(Arc<AtomicBool>);
impl Drop for DropFlag {
fn drop(&mut self) {
self.0.store(true, Ordering::SeqCst);
}
}
/// Abnormal death of one linked actor cooperatively stops its non-trapping
/// peer, and the peer's Drop guards run during the propagated unwind.
#[test]
fn linked_pair_one_panics_other_is_stopped() {
let dropped = Arc::new(AtomicBool::new(false));
let saw_stopped = Arc::new(AtomicBool::new(false));
let (d, s) = (dropped.clone(), saw_stopped.clone());
run(move || {
// B parks on recv (kept open by a retained sender) so it is alive and
// suspended when the propagated stop arrives.
let hb = spawn(move || {
let _g = DropFlag(d);
let (tx, rx) = channel::<u8>();
let _keep = tx;
let _ = rx.recv(); // parks until the link-propagated stop unwinds us
});
let b = hb.pid();
let down_b = monitor(b);
// A links B, then panics. The panic propagates along the link to B.
let ha = spawn(move || {
link(b);
panic!("boom");
});
let dn = down_b.recv().expect("monitor channel closed before Down");
assert_eq!(dn.pid, b, "Down reported the wrong pid");
if matches!(dn.reason, DownReason::Stopped) {
s.store(true, Ordering::SeqCst);
}
drop(ha);
let _ = hb.join();
});
assert!(
saw_stopped.load(Ordering::SeqCst),
"peer should be Stopped by the propagated abnormal exit"
);
assert!(
dropped.load(Ordering::SeqCst),
"peer's Drop guard must run during the propagated cancellation unwind"
);
}
/// A trapping peer receives an abnormal death as an `ExitSignal` message and
/// keeps running instead of being torn down.
#[test]
fn trap_exit_turns_abnormal_death_into_a_message() {
let ok = Arc::new(AtomicBool::new(false));
let o = ok.clone();
run(move || {
let h = spawn(move || {
let inbox = trap_exit();
// Spawn the peer from here so we control ordering: it is enqueued
// behind us and runs once we park on `inbox.recv()`.
let ha = spawn(|| panic!("peer down"));
let a = ha.pid();
link(a);
// Parks here; A then runs, panics, and the trap message wakes us.
let sig = inbox.recv().expect("trap inbox closed without a signal");
if sig.from == a && matches!(sig.reason, DownReason::Panic) {
o.store(true, Ordering::SeqCst);
}
drop(ha);
});
let _ = h.join();
});
assert!(
ok.load(Ordering::SeqCst),
"trapping peer should receive ExitSignal{{from, Panic}} and survive"
);
}
/// A normal exit never propagates — a trapping linked peer sees no message and
/// stays alive.
#[test]
fn normal_exit_does_not_propagate() {
let empty = Arc::new(AtomicBool::new(false));
let e = empty.clone();
run(move || {
let h = spawn(move || {
let inbox = trap_exit();
let ha = spawn(|| { /* exits normally, immediately */ });
let a = ha.pid();
link(a);
yield_now(); // let A run to completion and finalize
// A exited normally: nothing should have landed on the inbox.
if let Ok(None) = inbox.try_recv() {
e.store(true, Ordering::SeqCst);
}
drop(ha);
});
let _ = h.join();
});
assert!(
empty.load(Ordering::SeqCst),
"normal exit must not deliver any ExitSignal to a trapping peer"
);
}
/// Linking an already-dead pid from a non-trapping actor stops the caller
/// (no silent no-op).
#[test]
fn link_to_dead_pid_stops_a_nontrapping_caller() {
let dropped = Arc::new(AtomicBool::new(false));
let saw_stopped = Arc::new(AtomicBool::new(false));
let (d, s) = (dropped.clone(), saw_stopped.clone());
run(move || {
let ha = spawn(|| {});
let a = ha.pid();
ha.join().unwrap(); // A finishes and is reclaimed → `a` is now stale
let hb = spawn(move || {
let _g = DropFlag(d);
link(a); // dead target, not trapping → request_stop(self)
loop {
smarm::check!(); // observation point to realize the stop
}
});
let b = hb.pid();
let down_b = monitor(b);
let dn = down_b.recv().expect("monitor channel closed before Down");
if matches!(dn.reason, DownReason::Stopped) {
s.store(true, Ordering::SeqCst);
}
let _ = hb.join();
});
assert!(
saw_stopped.load(Ordering::SeqCst),
"linking a dead pid (non-trapping) should stop the caller"
);
assert!(dropped.load(Ordering::SeqCst), "Drop guard must run");
}
/// Linking an already-dead pid from a trapping actor delivers an immediate
/// `NoProc` exit message instead of killing it.
#[test]
fn link_to_dead_pid_messages_a_trapping_caller() {
let ok = Arc::new(AtomicBool::new(false));
let o = ok.clone();
run(move || {
let ha = spawn(|| {});
let a = ha.pid();
ha.join().unwrap(); // `a` is now stale
let hb = spawn(move || {
let inbox = trap_exit();
link(a); // dead target, trapping → immediate NoProc message
let sig = inbox.recv().expect("trap inbox closed without a signal");
if sig.from == a && matches!(sig.reason, DownReason::NoProc) {
o.store(true, Ordering::SeqCst);
}
});
let _ = hb.join();
});
assert!(
ok.load(Ordering::SeqCst),
"linking a dead pid (trapping) should deliver a NoProc ExitSignal"
);
}
/// `unlink` removes the relationship in both directions: an abnormal death no
/// longer reaches the formerly-linked peer.
#[test]
fn unlink_prevents_propagation() {
let survived = Arc::new(AtomicBool::new(false));
let sv = survived.clone();
run(move || {
let h = spawn(move || {
let b = self_pid();
let inbox = trap_exit();
let ha = spawn(move || {
link(b);
unlink(b);
panic!("boom"); // abnormal, but the link is gone
});
yield_now(); // let A link, unlink, and panic
// Unlinked before death → no ExitSignal should have arrived.
if let Ok(None) = inbox.try_recv() {
sv.store(true, Ordering::SeqCst);
}
drop(ha);
});
let _ = h.join();
});
assert!(
survived.load(Ordering::SeqCst),
"unlinked peer must not receive a propagated exit"
);
}