src/corosio/src/detail/select/op.hpp

74.2% Lines (95/128) 84.2% Functions (16/19) 66.7% Branches (24/36)
src/corosio/src/detail/select/op.hpp
Line Branch Hits Source Code
1 //
2 // Copyright (c) 2026 Steve Gerbino
3 //
4 // Distributed under the Boost Software License, Version 1.0. (See accompanying
5 // file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
6 //
7 // Official repository: https://github.com/cppalliance/corosio
8 //
9
10 #ifndef BOOST_COROSIO_DETAIL_SELECT_OP_HPP
11 #define BOOST_COROSIO_DETAIL_SELECT_OP_HPP
12
13 #include <boost/corosio/detail/platform.hpp>
14
15 #if BOOST_COROSIO_HAS_SELECT
16
17 #include <boost/corosio/detail/config.hpp>
18 #include <boost/corosio/io_object.hpp>
19 #include <boost/corosio/endpoint.hpp>
20 #include <boost/capy/ex/executor_ref.hpp>
21 #include <coroutine>
22 #include <boost/capy/error.hpp>
23 #include <system_error>
24
25 #include "src/detail/make_err.hpp"
26 #include "src/detail/dispatch_coro.hpp"
27 #include "src/detail/scheduler_op.hpp"
28 #include "src/detail/endpoint_convert.hpp"
29
30 #include <unistd.h>
31 #include <errno.h>
32 #include <fcntl.h>
33
34 #include <atomic>
35 #include <cstddef>
36 #include <memory>
37 #include <optional>
38 #include <stop_token>
39
40 #include <netinet/in.h>
41 #include <sys/select.h>
42 #include <sys/socket.h>
43 #include <sys/uio.h>
44
45 /*
46 select Operation State
47 ======================
48
49 Each async I/O operation has a corresponding select_op-derived struct that
50 holds the operation's state while it's in flight. The socket impl owns
51 fixed slots for each operation type (conn_, rd_, wr_), so only one
52 operation of each type can be pending per socket at a time.
53
54 This mirrors the epoll_op design for consistency across backends.
55
56 Completion vs Cancellation Race
57 -------------------------------
58 The `registered` atomic uses a tri-state (unregistered, registering,
59 registered) to handle two races: (1) between register_fd() and the
60 reactor seeing an event, and (2) between reactor completion and cancel().
61
62 The registering state closes the window where an event could arrive
63 after register_fd() but before the boolean was set. The reactor and
64 cancel() both treat registering the same as registered when claiming.
65
66 Whoever atomically exchanges to unregistered "claims" the operation
67 and is responsible for completing it. The loser sees unregistered and
68 does nothing. The initiating thread uses compare_exchange to transition
69 from registering to registered; if this fails, the reactor or cancel
70 already claimed the op.
71
72 Impl Lifetime Management
73 ------------------------
74 When cancel() posts an op to the scheduler's ready queue, the socket impl
75 might be destroyed before the scheduler processes the op. The `impl_ptr`
76 member holds a shared_ptr to the impl, keeping it alive until the op
77 completes.
78
79 EOF Detection
80 -------------
81 For reads, 0 bytes with no error means EOF. But an empty user buffer also
82 returns 0 bytes. The `empty_buffer_read` flag distinguishes these cases.
83
84 SIGPIPE Prevention
85 ------------------
86 Writes use sendmsg() with MSG_NOSIGNAL instead of writev() to prevent
87 SIGPIPE when the peer has closed.
88 */
89
90 namespace boost::corosio::detail {
91
92 // Forward declarations for cancellation support
93 class select_socket_impl;
94 class select_acceptor_impl;
95
96 /** Registration state for async operations.
97
98 Tri-state enum to handle the race between register_fd() and
99 run_reactor() seeing an event. Setting REGISTERING before
100 calling register_fd() ensures events delivered during the
101 registration window are not dropped.
102 */
103 enum class select_registration_state : std::uint8_t
104 {
105 unregistered, ///< Not registered with reactor
106 registering, ///< register_fd() called, not yet confirmed
107 registered ///< Fully registered, ready for events
108 };
109
110 struct select_op : scheduler_op
111 {
112 struct canceller
113 {
114 select_op* op;
115 void operator()() const noexcept;
116 };
117
118 std::coroutine_handle<> h;
119 capy::executor_ref ex;
120 std::error_code* ec_out = nullptr;
121 std::size_t* bytes_out = nullptr;
122
123 int fd = -1;
124 int errn = 0;
125 std::size_t bytes_transferred = 0;
126
127 std::atomic<bool> cancelled{false};
128 std::atomic<select_registration_state> registered{select_registration_state::unregistered};
129 std::optional<std::stop_callback<canceller>> stop_cb;
130
131 // Prevents use-after-free when socket is closed with pending ops.
132 std::shared_ptr<void> impl_ptr;
133
134 // For stop_token cancellation - pointer to owning socket/acceptor impl.
135 select_socket_impl* socket_impl_ = nullptr;
136 select_acceptor_impl* acceptor_impl_ = nullptr;
137
138 21399 select_op() = default;
139
140 250500 void reset() noexcept
141 {
142 250500 fd = -1;
143 250500 errn = 0;
144 250500 bytes_transferred = 0;
145 250500 cancelled.store(false, std::memory_order_relaxed);
146 250500 registered.store(select_registration_state::unregistered, std::memory_order_relaxed);
147 250500 impl_ptr.reset();
148 250500 socket_impl_ = nullptr;
149 250500 acceptor_impl_ = nullptr;
150 250500 }
151
152 243389 void operator()() override
153 {
154 243389 stop_cb.reset();
155
156
1/2
✓ Branch 0 taken 243389 times.
✗ Branch 1 not taken.
243389 if (ec_out)
157 {
158
2/2
✓ Branch 1 taken 201 times.
✓ Branch 2 taken 243188 times.
243389 if (cancelled.load(std::memory_order_acquire))
159 201 *ec_out = capy::error::canceled;
160
2/2
✓ Branch 0 taken 1 time.
✓ Branch 1 taken 243187 times.
243188 else if (errn != 0)
161 1 *ec_out = make_err(errn);
162
6/6
✓ Branch 1 taken 121579 times.
✓ Branch 2 taken 121608 times.
✓ Branch 3 taken 5 times.
✓ Branch 4 taken 121574 times.
✓ Branch 5 taken 5 times.
✓ Branch 6 taken 243182 times.
243187 else if (is_read_operation() && bytes_transferred == 0)
163 5 *ec_out = capy::error::eof;
164 else
165 243182 *ec_out = {};
166 }
167
168
1/2
✓ Branch 0 taken 243389 times.
✗ Branch 1 not taken.
243389 if (bytes_out)
169 243389 *bytes_out = bytes_transferred;
170
171 // Move to stack before destroying the frame
172 243389 capy::executor_ref saved_ex( std::move( ex ) );
173 243389 std::coroutine_handle<> saved_h( std::move( h ) );
174 243389 impl_ptr.reset();
175
2/2
✓ Branch 1 taken 243389 times.
✓ Branch 4 taken 243389 times.
243389 dispatch_coro(saved_ex, saved_h).resume();
176 243389 }
177
178 121607 virtual bool is_read_operation() const noexcept { return false; }
179 virtual void cancel() noexcept = 0;
180
181 void destroy() override
182 {
183 stop_cb.reset();
184 impl_ptr.reset();
185 }
186
187 32776 void request_cancel() noexcept
188 {
189 32776 cancelled.store(true, std::memory_order_release);
190 32776 }
191
192 void start(std::stop_token token)
193 {
194 cancelled.store(false, std::memory_order_release);
195 stop_cb.reset();
196 socket_impl_ = nullptr;
197 acceptor_impl_ = nullptr;
198
199 if (token.stop_possible())
200 stop_cb.emplace(token, canceller{this});
201 }
202
203 246944 void start(std::stop_token token, select_socket_impl* impl)
204 {
205 246944 cancelled.store(false, std::memory_order_release);
206 246944 stop_cb.reset();
207 246944 socket_impl_ = impl;
208 246944 acceptor_impl_ = nullptr;
209
210
2/2
✓ Branch 1 taken 100 times.
✓ Branch 2 taken 246844 times.
246944 if (token.stop_possible())
211 100 stop_cb.emplace(token, canceller{this});
212 246944 }
213
214 3556 void start(std::stop_token token, select_acceptor_impl* impl)
215 {
216 3556 cancelled.store(false, std::memory_order_release);
217 3556 stop_cb.reset();
218 3556 socket_impl_ = nullptr;
219 3556 acceptor_impl_ = impl;
220
221
1/2
✗ Branch 1 not taken.
✓ Branch 2 taken 3556 times.
3556 if (token.stop_possible())
222 stop_cb.emplace(token, canceller{this});
223 3556 }
224
225 250336 void complete(int err, std::size_t bytes) noexcept
226 {
227 250336 errn = err;
228 250336 bytes_transferred = bytes;
229 250336 }
230
231 virtual void perform_io() noexcept {}
232 };
233
234
235 struct select_connect_op : select_op
236 {
237 endpoint target_endpoint;
238
239 3555 void reset() noexcept
240 {
241 3555 select_op::reset();
242 3555 target_endpoint = endpoint{};
243 3555 }
244
245 3555 void perform_io() noexcept override
246 {
247 // connect() completion status is retrieved via SO_ERROR, not return value
248 3555 int err = 0;
249 3555 socklen_t len = sizeof(err);
250
1/2
✗ Branch 1 not taken.
✓ Branch 2 taken 3555 times.
3555 if (::getsockopt(fd, SOL_SOCKET, SO_ERROR, &err, &len) < 0)
251 err = errno;
252 3555 complete(err, 0);
253 3555 }
254
255 // Defined in sockets.cpp where select_socket_impl is complete
256 void operator()() override;
257 void cancel() noexcept override;
258 };
259
260
261 struct select_read_op : select_op
262 {
263 static constexpr std::size_t max_buffers = 16;
264 iovec iovecs[max_buffers];
265 int iovec_count = 0;
266 bool empty_buffer_read = false;
267
268 121580 bool is_read_operation() const noexcept override
269 {
270 121580 return !empty_buffer_read;
271 }
272
273 121777 void reset() noexcept
274 {
275 121777 select_op::reset();
276 121777 iovec_count = 0;
277 121777 empty_buffer_read = false;
278 121777 }
279
280 124 void perform_io() noexcept override
281 {
282 124 ssize_t n = ::readv(fd, iovecs, iovec_count);
283
1/2
✓ Branch 0 taken 124 times.
✗ Branch 1 not taken.
124 if (n >= 0)
284 124 complete(0, static_cast<std::size_t>(n));
285 else
286 complete(errno, 0);
287 124 }
288
289 void cancel() noexcept override;
290 };
291
292
293 struct select_write_op : select_op
294 {
295 static constexpr std::size_t max_buffers = 16;
296 iovec iovecs[max_buffers];
297 int iovec_count = 0;
298
299 121612 void reset() noexcept
300 {
301 121612 select_op::reset();
302 121612 iovec_count = 0;
303 121612 }
304
305 void perform_io() noexcept override
306 {
307 msghdr msg{};
308 msg.msg_iov = iovecs;
309 msg.msg_iovlen = static_cast<std::size_t>(iovec_count);
310
311 ssize_t n = ::sendmsg(fd, &msg, MSG_NOSIGNAL);
312 if (n >= 0)
313 complete(0, static_cast<std::size_t>(n));
314 else
315 complete(errno, 0);
316 }
317
318 void cancel() noexcept override;
319 };
320
321
322 struct select_accept_op : select_op
323 {
324 int accepted_fd = -1;
325 io_object::io_object_impl* peer_impl = nullptr;
326 io_object::io_object_impl** impl_out = nullptr;
327
328 3556 void reset() noexcept
329 {
330 3556 select_op::reset();
331 3556 accepted_fd = -1;
332 3556 peer_impl = nullptr;
333 3556 impl_out = nullptr;
334 3556 }
335
336 3551 void perform_io() noexcept override
337 {
338 3551 sockaddr_in addr{};
339 3551 socklen_t addrlen = sizeof(addr);
340
341 // Note: select backend uses accept() + fcntl instead of accept4()
342 // for broader POSIX compatibility
343 3551 int new_fd = ::accept(fd, reinterpret_cast<sockaddr*>(&addr), &addrlen);
344
345
1/2
✓ Branch 0 taken 3551 times.
✗ Branch 1 not taken.
3551 if (new_fd >= 0)
346 {
347 // Reject fds that exceed select()'s FD_SETSIZE limit.
348 // Better to fail now than during later async operations.
349
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 3551 times.
3551 if (new_fd >= FD_SETSIZE)
350 {
351 ::close(new_fd);
352 complete(EINVAL, 0);
353 return;
354 }
355
356 // Set non-blocking and close-on-exec flags.
357 // A non-blocking socket is essential for the async reactor;
358 // if we can't configure it, fail rather than risk blocking.
359 3551 int flags = ::fcntl(new_fd, F_GETFL, 0);
360
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 3551 times.
3551 if (flags == -1)
361 {
362 int err = errno;
363 ::close(new_fd);
364 complete(err, 0);
365 return;
366 }
367
368
1/2
✗ Branch 1 not taken.
✓ Branch 2 taken 3551 times.
3551 if (::fcntl(new_fd, F_SETFL, flags | O_NONBLOCK) == -1)
369 {
370 int err = errno;
371 ::close(new_fd);
372 complete(err, 0);
373 return;
374 }
375
376
1/2
✗ Branch 1 not taken.
✓ Branch 2 taken 3551 times.
3551 if (::fcntl(new_fd, F_SETFD, FD_CLOEXEC) == -1)
377 {
378 int err = errno;
379 ::close(new_fd);
380 complete(err, 0);
381 return;
382 }
383
384 3551 accepted_fd = new_fd;
385 3551 complete(0, 0);
386 }
387 else
388 {
389 complete(errno, 0);
390 }
391 }
392
393 // Defined in acceptors.cpp where select_acceptor_impl is complete
394 void operator()() override;
395 void cancel() noexcept override;
396 };
397
398 } // namespace boost::corosio::detail
399
400 #endif // BOOST_COROSIO_HAS_SELECT
401
402 #endif // BOOST_COROSIO_DETAIL_SELECT_OP_HPP
403