Fix bug where backpressure happen too early

This commit is contained in:
fakeshadow 2021-04-14 15:57:32 +08:00
parent e0fb67f646
commit 501488f86e
1 changed files with 98 additions and 25 deletions

View File

@ -83,9 +83,55 @@ struct Accept {
handles: Vec<WorkerHandleAccept>,
srv: Server,
next: usize,
avail: Availability,
backpressure: bool,
}
/// Array of u128 with every bit as marker for a worker handle's availability.
struct Availability([u128; 4]);
impl Default for Availability {
fn default() -> Self {
Self([0; 4])
}
}
impl Availability {
/// Check if any worker handle is available
fn available(&self) -> bool {
self.0.iter().any(|a| *a != 0)
}
/// Set worker handle available state by index.
fn set_available(&mut self, idx: usize, avail: bool) {
let (offset, idx) = if idx < 128 {
(0, idx)
} else if idx < 128 * 2 {
(1, idx - 128)
} else if idx < 128 * 3 {
(2, idx - 128 * 2)
} else if idx < 128 * 4 {
(3, idx - 128 * 3)
} else {
unreachable!("Max WorkerHandle count is 512")
};
if avail {
self.0[offset] |= 1 << idx as u128;
} else {
self.0[offset] ^= 1 << idx as u128;
}
}
/// Set all worker handle to available state.
/// This would result in a re-check on all workers' availability.
fn set_available_all(&mut self, handles: &[WorkerHandleAccept]) {
handles.iter().for_each(|handle| {
self.set_available(handle.idx, true);
})
}
}
/// This function defines errors that are per-connection. Which basically
/// means that if we get this error from `accept()` system call it means
/// next connection might be ready to be accepted.
@ -116,6 +162,7 @@ impl Accept {
System::set_current(sys);
let (mut accept, sockets) =
Accept::new_with_sockets(poll, waker, socks, handles, srv);
accept.poll_with(sockets);
})
.unwrap();
@ -148,16 +195,31 @@ impl Accept {
});
}
let accept = Accept {
let accept = Accept::new(poll, waker, handles, srv);
(accept, sockets)
}
fn new(
poll: Poll,
waker: WakerQueue,
handles: Vec<WorkerHandleAccept>,
srv: Server,
) -> Self {
let mut avail = Availability::default();
// Assume all handles are avail at construct time.
avail.set_available_all(&handles);
Self {
poll,
waker,
handles,
srv,
next: 0,
avail,
backpressure: false,
};
(accept, sockets)
}
}
fn poll_with(&mut self, mut sockets: Slab<ServerSocketInfo>) {
@ -190,6 +252,8 @@ impl Accept {
// from backpressure.
Some(WakerInterest::WorkerAvailable) => {
drop(guard);
// Assume all worker are avail as no worker index returned.
self.avail.set_available_all(&self.handles);
self.maybe_backpressure(&mut sockets, false);
}
// a new worker thread is made and it's handle would be added to Accept
@ -197,6 +261,7 @@ impl Accept {
drop(guard);
// maybe we want to recover from a backpressure.
self.maybe_backpressure(&mut sockets, false);
self.avail.set_available(handle.idx, true);
self.handles.push(handle);
}
// got timer interest and it's time to try register socket(s) again
@ -342,27 +407,25 @@ impl Accept {
if self.backpressure {
// send_connection would remove fault worker from handles.
// worst case here is conn get dropped after all handles are gone.
while !self.handles.is_empty() {
match self.send_connection(sockets, conn) {
Ok(_) => return,
Err(c) => conn = c,
}
while let Err(c) = self.send_connection(sockets, conn) {
conn = c
}
} else {
// Do one round and try to send conn to all workers until it succeed.
// Start from self.next.
let mut idx = 0;
while idx < self.handles.len() {
idx += 1;
if self.handles[self.next].available() {
while self.avail.available() {
let next = self.next;
let idx = self.handles[next].idx;
if self.handles[next].available() {
self.avail.set_available(idx, true);
match self.send_connection(sockets, conn) {
Ok(_) => return,
Err(c) => conn = c,
}
} else {
self.avail.set_available(idx, false);
self.set_next();
}
}
// Sending Conn failed due to either all workers are in error or not available.
// Enter backpressure state and try again.
self.maybe_backpressure(sockets, true);
@ -370,11 +433,6 @@ impl Accept {
}
}
// Set next worker handle that would accept work.
fn set_next(&mut self) {
self.next = (self.next + 1) % self.handles.len();
}
// Send connection to worker and handle error.
fn send_connection(
&mut self,
@ -387,11 +445,10 @@ impl Accept {
Ok(())
}
Err(conn) => {
// worker lost contact and could be gone. a message is sent to
// `ServerBuilder` future to notify it a new worker should be made.
// after that remove the fault worker and enter backpressure if necessary.
self.srv.worker_faulted(self.handles[self.next].idx);
self.handles.swap_remove(self.next);
// Worker thread is error and could be gone.
// Remove worker handle and notify `ServerBuilder`.
self.remove_next();
if self.handles.is_empty() {
error!("No workers");
self.maybe_backpressure(sockets, true);
@ -401,6 +458,7 @@ impl Accept {
} else if self.handles.len() <= self.next {
self.next = 0;
}
Err(conn)
}
}
@ -445,4 +503,19 @@ impl Accept {
};
}
}
/// Set next worker handle that would accept connection.
fn set_next(&mut self) {
self.next = (self.next + 1) % self.handles.len();
}
/// Remove next worker handle that fail to accept connection.
fn remove_next(&mut self) {
let handle = self.handles.swap_remove(self.next);
let idx = handle.idx;
// A message is sent to `ServerBuilder` future to notify it a new worker
// should be made.
self.srv.worker_faulted(idx);
self.avail.set_available(idx, false);
}
}