Driver: Fix scheduler crash after task closure

A removed audio task could still have one or more driver messages left in its queue, leading to a crash when the id->mixer lookup failed. This removes an unwrap which is invalid under these assumptions and includes an extra cleanup measure for message forwarders under the same circumstances.

This was tested using `cargo make ready`.
This commit is contained in:
Kyle Simpson
2023-06-07 21:15:30 +01:00
parent c976d50cc5
commit 77e3916bdc
4 changed files with 43 additions and 19 deletions

View File

@@ -75,6 +75,8 @@ pub struct ParkedMixer {
/// The last known cost of executing this task, if it had to be moved
/// due to a limit on thread resources.
pub last_cost: Option<Duration>,
/// Handle to any forwarder task, used if this mixer is culled while idle.
pub cull_handle: Option<Sender<()>>,
}
#[allow(missing_docs)]
@@ -88,6 +90,7 @@ impl ParkedMixer {
rtp_timestamp: random::<u32>(),
park_time: Instant::now(),
last_cost: None,
cull_handle: None,
}
}
@@ -95,14 +98,28 @@ impl ParkedMixer {
///
/// Any requests which would cause this mixer to become live will terminate
/// this task.
pub fn spawn_forwarder(&self, tx: Sender<SchedulerMessage>, id: TaskId) {
pub fn spawn_forwarder(&mut self, tx: Sender<SchedulerMessage>, id: TaskId) {
let (kill_tx, kill_rx) = flume::bounded(1);
self.cull_handle = Some(kill_tx);
let remote_rx = self.mixer.mix_rx.clone();
tokio::spawn(async move {
while let Ok(msg) = remote_rx.recv_async().await {
let exit = msg.is_mixer_now_live();
let dead = tx.send_async(SchedulerMessage::Do(id, msg)).await.is_err();
if exit || dead {
break;
loop {
tokio::select! {
biased;
_ = kill_rx.recv_async() => break,
msg = remote_rx.recv_async() => {
let exit = if let Ok(msg) = msg {
let remove_self = msg.is_mixer_now_live();
tx.send_async(SchedulerMessage::Do(id, msg)).await.is_err() || remove_self
} else {
true
};
if exit {
break;
}
}
}
}
});