From eb882a69b6c5b841b19e05494cfe0f09289b6af1 Mon Sep 17 00:00:00 2001 From: Matt Corallo Date: Mon, 28 Aug 2023 18:39:04 +0000 Subject: [PATCH] Drop dep `tokio`'s `io-util` feat as it broke MSRV and isn't useful We use `tokio`'s `io-util` feature to provide the `Async{Read,Write}Ext` traits, which allow us to simply launch a read future or `poll_write` directly as well as `split` the `TcpStream` into a read/write half. However, these traits aren't actually doing much for us - they are really just wrapping the `readable` future (which we can trivially use ourselves) and `poll_write` isn't doing anything for us that `poll_write_ready` can't. Similarly, the split logic is actually just `Arc`ing the `TcpStream` and busy-waiting when an operation is busy to prevent concurrent reads/writes. However, there's no reason to prevent concurrent access at the stream level - we aren't ever concurrently writing or reading (though we may concurrently read and write, which is fine). Worse, the `io-util` feature broke MSRV (though they're likely to fix this upstream) and carries two additional dependencies (only one on the latest upstream tokio). Thus, we simply drop the dependency here. Fixes #2527. --- lightning-net-tokio/Cargo.toml | 4 +- lightning-net-tokio/src/lib.rs | 75 ++++++++++++++++------------------ 2 files changed, 37 insertions(+), 42 deletions(-) diff --git a/lightning-net-tokio/Cargo.toml b/lightning-net-tokio/Cargo.toml index c2befdae5..d2ac6e547 100644 --- a/lightning-net-tokio/Cargo.toml +++ b/lightning-net-tokio/Cargo.toml @@ -17,8 +17,8 @@ rustdoc-args = ["--cfg", "docsrs"] [dependencies] bitcoin = "0.29.0" lightning = { version = "0.0.116", path = "../lightning" } -tokio = { version = "1.0", features = [ "io-util", "rt", "sync", "net", "time" ] } +tokio = { version = "1.0", features = [ "rt", "sync", "net", "time" ] } [dev-dependencies] -tokio = { version = "1.14", features = [ "io-util", "macros", "rt", "rt-multi-thread", "sync", "net", "time" ] } +tokio = { version = "1.14", features = [ "macros", "rt", "rt-multi-thread", "sync", "net", "time" ] } lightning = { version = "0.0.116", path = "../lightning", features = ["_test_utils"] } diff --git a/lightning-net-tokio/src/lib.rs b/lightning-net-tokio/src/lib.rs index d6d800416..6e2ea3f14 100644 --- a/lightning-net-tokio/src/lib.rs +++ b/lightning-net-tokio/src/lib.rs @@ -31,10 +31,10 @@ use bitcoin::secp256k1::PublicKey; -use tokio::net::TcpStream; +use tokio::net::{tcp, TcpStream}; use tokio::{io, time}; use tokio::sync::mpsc; -use tokio::io::{AsyncReadExt, AsyncWrite, AsyncWriteExt}; +use tokio::io::AsyncWrite; use lightning::ln::peer_handler; use lightning::ln::peer_handler::SocketDescriptor as LnSocketTrait; @@ -59,7 +59,7 @@ static ID_COUNTER: AtomicU64 = AtomicU64::new(0); // define a trivial two- and three- select macro with the specific types we need and just use that. pub(crate) enum SelectorOutput { - A(Option<()>), B(Option<()>), C(tokio::io::Result), + A(Option<()>), B(Option<()>), C(tokio::io::Result<()>), } pub(crate) struct TwoSelector< @@ -87,7 +87,7 @@ impl< } pub(crate) struct ThreeSelector< - A: Future> + Unpin, B: Future> + Unpin, C: Future> + Unpin + A: Future> + Unpin, B: Future> + Unpin, C: Future> + Unpin > { pub a: A, pub b: B, @@ -95,7 +95,7 @@ pub(crate) struct ThreeSelector< } impl< - A: Future> + Unpin, B: Future> + Unpin, C: Future> + Unpin + A: Future> + Unpin, B: Future> + Unpin, C: Future> + Unpin > Future for ThreeSelector { type Output = SelectorOutput; fn poll(mut self: Pin<&mut Self>, ctx: &mut task::Context<'_>) -> Poll { @@ -119,7 +119,7 @@ impl< /// Connection object (in an Arc>) in each SocketDescriptor we create as well as in the /// read future (which is returned by schedule_read). struct Connection { - writer: Option>, + writer: Option>, // Because our PeerManager is templated by user-provided types, and we can't (as far as I can // tell) have a const RawWakerVTable built out of templated functions, we need some indirection // between being woken up with write-ready and calling PeerManager::write_buffer_space_avail. @@ -156,7 +156,7 @@ impl Connection { async fn schedule_read( peer_manager: PM, us: Arc>, - mut reader: io::ReadHalf, + reader: Arc, mut read_wake_receiver: mpsc::Receiver<()>, mut write_avail_receiver: mpsc::Receiver<()>, ) where PM::Target: APeerManager { @@ -200,7 +200,7 @@ impl Connection { ThreeSelector { a: Box::pin(write_avail_receiver.recv()), b: Box::pin(read_wake_receiver.recv()), - c: Box::pin(reader.read(&mut buf)), + c: Box::pin(reader.readable()), }.await }; match select_result { @@ -211,8 +211,9 @@ impl Connection { } }, SelectorOutput::B(_) => {}, - SelectorOutput::C(read) => { - match read { + SelectorOutput::C(res) => { + if res.is_err() { break Disconnect::PeerDisconnected; } + match reader.try_read(&mut buf) { Ok(0) => break Disconnect::PeerDisconnected, Ok(len) => { let read_res = peer_manager.as_ref().read_event(&mut our_descriptor, &buf[0..len]); @@ -226,7 +227,11 @@ impl Connection { Err(_) => break Disconnect::CloseConnection, } }, - Err(_) => break Disconnect::PeerDisconnected, + Err(e) if e.kind() == std::io::ErrorKind::WouldBlock => { + // readable() is allowed to spuriously wake, so we have to handle + // WouldBlock here. + }, + Err(e) => break Disconnect::PeerDisconnected, } }, } @@ -239,18 +244,14 @@ impl Connection { // here. let _ = tokio::task::yield_now().await; }; - let writer_option = us.lock().unwrap().writer.take(); - if let Some(mut writer) = writer_option { - // If the socket is already closed, shutdown() will fail, so just ignore it. - let _ = writer.shutdown().await; - } + us.lock().unwrap().writer.take(); if let Disconnect::PeerDisconnected = disconnect_type { peer_manager.as_ref().socket_disconnected(&our_descriptor); peer_manager.as_ref().process_events(); } } - fn new(stream: StdTcpStream) -> (io::ReadHalf, mpsc::Receiver<()>, mpsc::Receiver<()>, Arc>) { + fn new(stream: StdTcpStream) -> (Arc, mpsc::Receiver<()>, mpsc::Receiver<()>, Arc>) { // We only ever need a channel of depth 1 here: if we returned a non-full write to the // PeerManager, we will eventually get notified that there is room in the socket to write // new bytes, which will generate an event. That event will be popped off the queue before @@ -262,11 +263,11 @@ impl Connection { // false. let (read_waker, read_receiver) = mpsc::channel(1); stream.set_nonblocking(true).unwrap(); - let (reader, writer) = io::split(TcpStream::from_std(stream).unwrap()); + let tokio_stream = Arc::new(TcpStream::from_std(stream).unwrap()); - (reader, write_receiver, read_receiver, + (Arc::clone(&tokio_stream), write_receiver, read_receiver, Arc::new(Mutex::new(Self { - writer: Some(writer), write_avail, read_waker, read_paused: false, + writer: Some(tokio_stream), write_avail, read_waker, read_paused: false, rl_requested_disconnect: false, id: ID_COUNTER.fetch_add(1, Ordering::AcqRel) }))) @@ -462,9 +463,9 @@ impl SocketDescriptor { } impl peer_handler::SocketDescriptor for SocketDescriptor { fn send_data(&mut self, data: &[u8], resume_read: bool) -> usize { - // To send data, we take a lock on our Connection to access the WriteHalf of the TcpStream, - // writing to it if there's room in the kernel buffer, or otherwise create a new Waker with - // a SocketDescriptor in it which can wake up the write_avail Sender, waking up the + // To send data, we take a lock on our Connection to access the TcpStream, writing to it if + // there's room in the kernel buffer, or otherwise create a new Waker with a + // SocketDescriptor in it which can wake up the write_avail Sender, waking up the // processing future which will call write_buffer_space_avail and we'll end up back here. let mut us = self.conn.lock().unwrap(); if us.writer.is_none() { @@ -484,24 +485,18 @@ impl peer_handler::SocketDescriptor for SocketDescriptor { let mut ctx = task::Context::from_waker(&waker); let mut written_len = 0; loop { - match std::pin::Pin::new(us.writer.as_mut().unwrap()).poll_write(&mut ctx, &data[written_len..]) { - task::Poll::Ready(Ok(res)) => { - // The tokio docs *seem* to indicate this can't happen, and I certainly don't - // know how to handle it if it does (cause it should be a Poll::Pending - // instead): - assert_ne!(res, 0); - written_len += res; - if written_len == data.len() { return written_len; } - }, - task::Poll::Ready(Err(e)) => { - // The tokio docs *seem* to indicate this can't happen, and I certainly don't - // know how to handle it if it does (cause it should be a Poll::Pending - // instead): - assert_ne!(e.kind(), io::ErrorKind::WouldBlock); - // Probably we've already been closed, just return what we have and let the - // read thread handle closing logic. - return written_len; + match us.writer.as_ref().unwrap().poll_write_ready(&mut ctx) { + task::Poll::Ready(Ok(())) => { + match us.writer.as_ref().unwrap().try_write(&data[written_len..]) { + Ok(res) => { + debug_assert_ne!(res, 0); + written_len += res; + if written_len == data.len() { return written_len; } + }, + Err(e) => return written_len, + } }, + task::Poll::Ready(Err(e)) => return written_len, task::Poll::Pending => { // We're queued up for a write event now, but we need to make sure we also // pause read given we're now waiting on the remote end to ACK (and in -- 2.39.5