Replace tokio-uring with io-uring (#993)

This is a fairly major change to swap out `tokio-uring` with the lower level `io-uring`, which has some upsides and downsides.

### Upsides

In `tokio-uring`, _every_ udp [recv_from](https://github.com/tokio-rs/tokio-uring/blob/7761222aa7f4bd48c559ca82e9535d47aac96d53/src/io/recv_from.rs#L20-L54) and [send_to](https://github.com/tokio-rs/tokio-uring/blob/7761222aa7f4bd48c559ca82e9535d47aac96d53/src/io/send_to.rs#L22-L70) performs **3** heap allocations (maybe even more in other parts of the code?) which is extremely wasteful in the context of a proxy that can be sending and receiving many thousands of packets a second. Moving to `io-uring` means we need to take responsibility for the lifetimes of memory being written/read by the kernel during I/O, but means we can minimize/get rid of memory allocations since we have the full context. For example, the QCMP loop now doesn't use the heap at all in favor of just reusing stack allocations.

Additionally, the current code which forwards packets either [downstream](https://github.com/googleforgames/quilkin/blob/ee0b70f654f4c01b1ce027985b80d7d1133c7432/src/components/proxy/packet_router.rs#L67-L110) or [upstream](https://github.com/googleforgames/quilkin/blob/ee0b70f654f4c01b1ce027985b80d7d1133c7432/src/components/proxy/sessions.rs#L166-L202) only ever sends 1 packet at a time per worker/session, the new code takes advantage of not being async/await by just sending up to a few thousand packets concurrently, reducing a (probably minor) throughput bottleneck.

### Downsides

A lot more code, some of which is unsafe, though slightly less than it could have been, as now the session and packet_router both share the same implementation. The non-linux code is also now separated since they are no longer really compatible since the io uring loop is not async so we can't pretend the code is the same between linux and non-linux, which also contributes to code increase.

Overall, it's just simply more complicated relative to the old code, but does give us tighter control.

Cargo.toml

@@ -38,6 +38,9 @@ categories = ["game-development", "network-programming"]
 edition.workspace = true
 exclude = ["docs", "build", "examples", "image"]
 
+[lints]
+workspace = true
+
 [[bench]]
 name = "read_write"
 harness = false
@@ -144,8 +147,10 @@ version = "0.1"
 features = ["client", "client-legacy"]
 
 [target.'cfg(target_os = "linux")'.dependencies]
+io-uring = { version = "0.6", default-features = false }
+libc = "0.2"
+slab = "0.4"
 sys-info = "0.9.1"
-tokio-uring = { version = "0.5", features = ["bytes"] }
 pprof = { version = "0.13.0", features = ["prost", "prost-codec"] }
 
 [dev-dependencies]
@@ -225,3 +230,6 @@ fixedstr = { version = "0.5", features = ["flex-str"] }
 parking_lot = "0.12.1"
 schemars = { version = "0.8.15", features = ["bytes", "url"] }
 url = { version = "2.4.1", features = ["serde"] }
+
+[workspace.lints.clippy]
+undocumented_unsafe_blocks = "deny"

crates/agones/Cargo.toml

@@ -23,6 +23,9 @@ license = "Apache-2.0"
 description = "End to end integration tests to be run against a Kubernetes cluster with Agones installed"
 readme = "README.md"
 
+[lints]
+workspace = true
+
 [dependencies]
 base64.workspace = true
 futures.workspace = true

crates/macros/Cargo.toml

@@ -30,7 +30,8 @@ edition = "2018"
 [lib]
 proc-macro = true
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+[lints]
+workspace = true
 
 [dependencies]
 proc-macro2 = "1.0.58"

crates/quilkin-proto/Cargo.toml

@@ -20,7 +20,8 @@ version = "0.1.0"
 edition.workspace = true
 license.workspace = true
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+[lints]
+workspace = true
 
 [dependencies]
 prost.workspace = true

crates/test/Cargo.toml

@@ -20,6 +20,9 @@ version = "0.1.0"
 edition = "2021"
 publish = false
 
+[lints]
+workspace = true
+
 [dependencies]
 async-channel.workspace = true
 once_cell.workspace = true

crates/test/tests/proxy.rs

@@ -124,17 +124,17 @@ trace_test!(uring_receiver, {
             config,
             tx,
             BUFFER_POOL.clone(),
-            shutdown_rx,
+            shutdown_rx.clone(),
         ),
     }
-    .spawn()
+    .spawn(shutdown_rx)
     .await
     .expect("failed to spawn task");
 
     // Drop the socket, otherwise it can
     drop(ws);
 
-    sb.timeout(500, ready.notified()).await;
+    let _ = sb.timeout(500, ready).await;
 
     let msg = "hello-downstream";
     tracing::debug!("sending packet");
@@ -166,7 +166,7 @@ trace_test!(
             config.clone(),
             tx,
             BUFFER_POOL.clone(),
-            shutdown_rx,
+            shutdown_rx.clone(),
         );
 
         const WORKER_COUNT: usize = 3;
@@ -179,12 +179,13 @@ trace_test!(
             &sessions,
             rx,
             BUFFER_POOL.clone(),
+            shutdown_rx,
         )
         .await
         .unwrap();
 
         for wn in workers {
-            sb.timeout(200, wn.notified()).await;
+            let _ = sb.timeout(200, wn).await;
         }
 
         let socket = std::sync::Arc::new(sb.client());