1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
use crate::execute_delay;
use crate::options::Options;
use crate::pretty::PrettyPrint;
use crate::response::build_response;
use crate::PORT_COUNTERS;
use crate::REQUESTS;
use crate::REQUEST_BYTES;
use crate::RESPONSES;
use crate::RESPONSE_BYTES;
use anyhow::Result;
use futures::stream::{select_all, unfold, StreamExt};
use futures::{AsyncReadExt, AsyncWriteExt};
use hyper::Response;
use owo_colors::OwoColorize;
use std::mem::MaybeUninit;
use std::net::SocketAddr;
use std::os::unix::io::{FromRawFd, IntoRawFd};
use std::sync::Arc;
use std::time::Duration;
use tracing::error;
use crate::create_listener;
use crate::ServerConfig;
pub fn run_thread(
id: usize,
addrs: Vec<SocketAddr>,
config: Arc<ServerConfig>,
opts: &Options,
) -> Result<()> {
use glommio::LocalExecutorBuilder;
use tracing::info;
let delay = opts.delay;
let meter = opts.meter;
let tcp_nodelay = opts.tcp_nodelay;
let verbose = opts.verbose;
let cpu_set =
core_affinity::get_core_ids().ok_or(anyhow::anyhow!("Failed to get thread affinity"))?;
let num_cpus = cpu_set.len();
let cpu_core = cpu_set.into_iter().nth(id % num_cpus).unwrap();
let num_entries = opts.uring_entries.next_power_of_two();
// Create multiple sockets manually with SO_REUSEPORT enabled before moving into the closure
let mut raw_fds = Vec::new();
for addr in &addrs {
let std_listener = create_listener(*addr, opts)?;
raw_fds.push(std_listener.into_raw_fd());
}
let builder = LocalExecutorBuilder::new(glommio::Placement::Fixed(cpu_core.id))
.ring_depth(num_entries as usize);
let handle = builder
.name(&format!("glommio-{}", id))
.spawn(move || async move {
// Create glommio listeners from raw fds
let listeners: Vec<glommio::net::TcpListener> = raw_fds
.into_iter()
.map(|fd| unsafe { glommio::net::TcpListener::from_raw_fd(fd) })
.collect();
// Combine all listeners into a single stream
let mut all_listeners =
select_all(listeners.into_iter().map(|l: glommio::net::TcpListener| {
let port = l.local_addr().unwrap().port();
Box::pin(unfold(
l,
move |listener: glommio::net::TcpListener| async move {
match listener.accept().await {
Ok(stream) => Some((Ok((stream, port)), listener)),
Err(e) => Some((Err(e), listener)),
}
},
))
}));
info!(
"Thread {} listening on {:?} (glommio cpu-{})",
id, addrs, cpu_core.id
);
loop {
let (stream, port) = match all_listeners.next().await {
Some(Ok((s, port))) => (s, port),
Some(Err(e)) => {
error!("Thread {} accept error: {}", id, e);
continue;
}
None => {
error!("Thread {} all listeners closed", id);
break;
}
};
// Enable TCP_NODELAY for lower latency
if tcp_nodelay {
if let Err(e) = stream.set_nodelay(true) {
error!("Failed to set TCP_NODELAY: {}", e);
}
}
let config = config.clone();
// Spawn task to handle the connection with glommio
glommio::spawn_local(async move {
if let Err(e) = handle_connection_glommio(
stream, port, config, false, meter, delay, verbose,
)
.await
{
error!("Error handling glommio connection: {}", e);
}
})
.detach();
}
})
.map_err(|e| anyhow::anyhow!("Failed to spawn glommio executor: {:?}", e))?;
handle.join().unwrap();
Ok(())
}
async fn handle_connection_glommio(
mut stream: glommio::net::TcpStream,
port: u16,
config: Arc<ServerConfig>,
http2: bool,
meter: bool,
delay: Option<Duration>,
verbose: u8,
) -> Result<usize> {
use http_wire::WireDecode;
if http2 {
return Err(anyhow::anyhow!("HTTP/2 not supported with glommio"));
}
let response_bytes = build_response(&config)?;
// Single buffer strategy for maximum performance:
// - In the common case (complete request in one read), we parse directly with zero copies
// - Only compact when we have leftover data from incomplete requests (rare case)
// - Glommio's read() accepts a slice, so we can read directly into spare capacity without unsafe tricks
let mut buf: Vec<u8> = Vec::with_capacity(8192);
let mut parsed = 0; // Number of bytes already parsed/consumed
let mut requests_served = 0;
loop {
// FAST PATH: If we've consumed all data, reset the buffer (zero-cost operation)
if parsed == buf.len() && parsed > 0 {
unsafe {
buf.set_len(0);
}
parsed = 0;
}
// SLOW PATH: If we have many dead bytes at the beginning, compact the buffer
// This only happens when requests are split across multiple reads (rare)
else if parsed > 4096 {
buf.copy_within(parsed.., 0);
unsafe {
buf.set_len(buf.len() - parsed);
}
parsed = 0;
}
// Ensure we have enough spare capacity for reading
// Only reserve if needed to avoid unnecessary allocations in the common case
let current_len = buf.len();
if buf.capacity() - current_len < 4096 {
buf.reserve(4096);
}
// Read directly into the buffer's spare capacity using a mutable slice
// Glommio accepts &mut [u8], which is simpler than monoio/tokio_uring's ownership model
// We temporarily extend the buffer to its capacity, read into the spare portion, then adjust
let capacity = buf.capacity();
unsafe {
buf.set_len(capacity);
}
let n = match stream.read(&mut buf[current_len..]).await {
Ok(0) => break, // Connection closed (EOF)
Ok(n) => n,
Err(e) => {
unsafe {
buf.set_len(current_len);
}
return Err(anyhow::Error::new(e));
}
};
// Update the buffer's length to include only the newly read data
unsafe {
buf.set_len(current_len + n);
}
// Parse all complete requests in the buffer
// This handles HTTP pipelining: multiple requests in a single read
let mut headers = [const { MaybeUninit::uninit() }; 128];
loop {
match http_wire::request::FullRequest::decode_uninit(&buf[parsed..], &mut headers) {
Ok((req, req_len)) => {
requests_served += 1;
parsed += req_len;
if verbose > 0 {
println!("↩ {}:\n{}", "request".bold(), req.pretty(verbose));
}
if meter {
REQUESTS.add(1);
REQUEST_BYTES.add(req_len);
let entry = PORT_COUNTERS.entry(port).or_default();
entry.requests.add(1);
entry.request_bytes.add(req_len);
}
if let Some(d) = delay {
execute_delay(d).await;
}
if verbose > 0 {
let mut print_builder = Response::builder().status(config.status);
for (k, v) in &config.headers {
print_builder = print_builder.header(k, v);
}
let print_resp = print_builder.body(config.body.clone()).unwrap();
println!("↪ {}:\n{}", "response".bold(), print_resp.pretty(verbose));
}
// Write the pre-encoded response
stream.write_all(&response_bytes).await?;
if meter {
RESPONSES.add(1);
RESPONSE_BYTES.add(response_bytes.len());
let entry = PORT_COUNTERS.entry(port).or_default();
entry.responses.add(1);
entry.response_bytes.add(response_bytes.len());
}
}
Err(_) => break, // Incomplete request or end of batch
}
}
// Loop continues: common case is parsed == buf.len(), so fast path activates on next iteration
}
Ok(requests_served)
}