include/boost/corosio/native/detail/epoll/epoll_scheduler.hpp

99.3% Lines (152/0/153) 100.0% List of functions (11/0/11)

[ ] prev/next file n p prev/next uncovered line

epoll_scheduler.hpp

f(x) Functions (11)

Function Calls Lines Blocks

Line	Hits	Source Code
1		//
2		// Copyright (c) 2026 Steve Gerbino
3		// Copyright (c) 2026 Michael Vandeberg
4		//
5		// Distributed under the Boost Software License, Version 1.0. (See accompanying
6		// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
7		//
8		// Official repository: https://github.com/cppalliance/corosio
9		//
10
11		#ifndef BOOST_COROSIO_NATIVE_DETAIL_EPOLL_EPOLL_SCHEDULER_HPP
12		#define BOOST_COROSIO_NATIVE_DETAIL_EPOLL_EPOLL_SCHEDULER_HPP
13
14		#include <boost/corosio/detail/platform.hpp>
15
16		#if BOOST_COROSIO_HAS_EPOLL
17
18		#include <boost/corosio/detail/config.hpp>
19		#include <boost/capy/ex/execution_context.hpp>
20
21		#include <boost/corosio/native/detail/reactor/reactor_scheduler.hpp>
22		#include <boost/corosio/native/detail/reactor/reactor_signal_pipe.hpp>
23
24		#include <boost/corosio/native/detail/epoll/epoll_traits.hpp>
25		#include <boost/corosio/detail/timer_service.hpp>
26		#include <boost/corosio/native/detail/make_err.hpp>
27		#include <boost/corosio/native/detail/posix/posix_resolver_service.hpp>
28		#include <boost/corosio/native/detail/posix/posix_signal_service.hpp>
29		#include <boost/corosio/native/detail/posix/posix_stream_file_service.hpp>
30		#include <boost/corosio/native/detail/posix/posix_random_access_file_service.hpp>
31
32		#include <boost/corosio/detail/except.hpp>
33
34		#include <atomic>
35		#include <chrono>
36		#include <cstdint>
37		#include <mutex>
38		#include <vector>
39
40		#include <errno.h>
41		#include <sys/epoll.h>
42		#include <sys/eventfd.h>
43		#include <sys/timerfd.h>
44		#include <unistd.h>
45
46		namespace boost::corosio::detail {
47
48		/** Linux scheduler using epoll for I/O multiplexing.
49
50		This scheduler implements the scheduler interface using Linux epoll
51		for efficient I/O event notification. It uses a single reactor model
52		where one thread runs epoll_wait while other threads
53		wait on a condition variable for handler work. This design provides:
54
55		- Handler parallelism: N posted handlers can execute on N threads
56		- No thundering herd: condition_variable wakes exactly one thread
57		- IOCP parity: Behavior matches Windows I/O completion port semantics
58
59		When threads call run(), they first try to execute queued handlers.
60		If the queue is empty and no reactor is running, one thread becomes
61		the reactor and runs epoll_wait. Other threads wait on a condition
62		variable until handlers are available.
63
64		@par Thread Safety
65		All public member functions are thread-safe.
66		*/
67		class BOOST_COROSIO_DECL epoll_scheduler final : public reactor_scheduler
68		{
69		public:
70		/** Construct the scheduler.
71
72		Creates an epoll instance, eventfd for reactor interruption,
73		and timerfd for kernel-managed timer expiry.
74
75		@param ctx Reference to the owning execution_context.
76		@param concurrency_hint Hint for expected thread count (unused).
77		*/
78		epoll_scheduler(capy::execution_context& ctx, int concurrency_hint = -1);
79
80		/// Destroy the scheduler.
81		~epoll_scheduler() override;
82
83		epoll_scheduler(epoll_scheduler const&) = delete;
84		epoll_scheduler& operator=(epoll_scheduler const&) = delete;
85
86		/// Shut down the scheduler, draining pending operations.
87		void shutdown() override;
88
89		/// Apply runtime configuration, resizing the event buffer.
90		void configure_reactor(
91		unsigned max_events,
92		unsigned budget_init,
93		unsigned budget_max,
94		unsigned unassisted) override;
95
96		/** Return the epoll file descriptor.
97
98		Used by socket services to register file descriptors
99		for I/O event notification.
100
101		@return The epoll file descriptor.
102		*/
103		int epoll_fd() const noexcept
104		{
105		return epoll_fd_;
106		}
107
108		/** Register a descriptor for persistent monitoring.
109
110		The fd is registered once and stays registered until explicitly
111		deregistered. Events are dispatched via reactor_descriptor_state which
112		tracks pending read/write/connect operations.
113
114		@param fd The file descriptor to register.
115		@param desc Pointer to descriptor data (stored in epoll_event.data.ptr).
116
117		@return The error if registration fails, otherwise a default
118		constructed error code.
119		*/
120		std::error_code
121		register_descriptor(int fd, reactor_descriptor_state* desc) const;
122
123		/** Deregister a persistently registered descriptor.
124
125		@param fd The file descriptor to deregister.
126		*/
127		void deregister_descriptor(int fd) const;
128
129		/// Watch the read end of the POSIX signal self-pipe (see scheduler.hpp).
130		[[nodiscard]] std::error_code
131	61x	register_signal_reader(int read_fd) override
132		{
133	61x	return register_descriptor(read_fd, signal_pipe_reader_.arm());
134		}
135
136		private:
137		void
138		run_task(lock_type& lock, context_type* ctx,
139		long timeout_us) override;
140		void interrupt_reactor() const override;
141		void update_timerfd() const;
142
143		int epoll_fd_;
144		int event_fd_;
145		int timer_fd_;
146
147		// Watches the global signal self-pipe's read end (armed lazily by
148		// register_signal_reader on the first signal registration).
149		reactor_signal_pipe_reader signal_pipe_reader_;
150
151		// Edge-triggered eventfd state
152		mutable std::atomic<bool> eventfd_armed_{false};
153
154		// Set when the earliest timer changes; flushed before epoll_wait
155		mutable std::atomic<bool> timerfd_stale_{false};
156
157		// Event buffer sized from max_events_per_poll_ (set at construction,
158		// resized by configure_reactor via io_context_options).
159		std::vector<epoll_event> event_buffer_;
160		};
161
162	1014x	inline epoll_scheduler::epoll_scheduler(capy::execution_context& ctx, int)
163	1014x	: epoll_fd_(-1)
164	1014x	, event_fd_(-1)
165	1014x	, timer_fd_(-1)
166	2028x	, event_buffer_(max_events_per_poll_)
167		{
168	1014x	epoll_fd_ = ::epoll_create1(EPOLL_CLOEXEC);
169	1014x	if (epoll_fd_ < 0)
170	1x	detail::throw_system_error(make_err(errno), "epoll_create1");
171
172	1013x	event_fd_ = ::eventfd(0, EFD_NONBLOCK \| EFD_CLOEXEC);
173	1013x	if (event_fd_ < 0)
174		{
175	1x	int errn = errno;
176	1x	::close(epoll_fd_);
177	1x	detail::throw_system_error(make_err(errn), "eventfd");
178		}
179
180	1012x	timer_fd_ = ::timerfd_create(CLOCK_MONOTONIC, TFD_NONBLOCK \| TFD_CLOEXEC);
181	1012x	if (timer_fd_ < 0)
182		{
183	1x	int errn = errno;
184	1x	::close(event_fd_);
185	1x	::close(epoll_fd_);
186	1x	detail::throw_system_error(make_err(errn), "timerfd_create");
187		}
188
189	1011x	epoll_event ev{};
190	1011x	ev.events = EPOLLIN \| EPOLLET;
191	1011x	ev.data.ptr = nullptr;
192	1011x	if (::epoll_ctl(epoll_fd_, EPOLL_CTL_ADD, event_fd_, &ev) < 0)
193		{
194	1x	int errn = errno;
195	1x	::close(timer_fd_);
196	1x	::close(event_fd_);
197	1x	::close(epoll_fd_);
198	1x	detail::throw_system_error(make_err(errn), "epoll_ctl");
199		}
200
201	1010x	epoll_event timer_ev{};
202	1010x	timer_ev.events = EPOLLIN \| EPOLLERR;
203	1010x	timer_ev.data.ptr = &timer_fd_;
204	1010x	if (::epoll_ctl(epoll_fd_, EPOLL_CTL_ADD, timer_fd_, &timer_ev) < 0)
205		{
206	1x	int errn = errno;
207	1x	::close(timer_fd_);
208	1x	::close(event_fd_);
209	1x	::close(epoll_fd_);
210	1x	detail::throw_system_error(make_err(errn), "epoll_ctl (timerfd)");
211		}
212
213	1009x	timer_svc_ = &get_timer_service(ctx, *this);
214	1009x	timer_svc_->set_on_earliest_changed(
215	5026x	timer_service::callback(this, [](void* p) {
216	4017x	auto* self = static_cast<epoll_scheduler*>(p);
217	4017x	self->timerfd_stale_.store(true, std::memory_order_release);
218	4017x	self->interrupt_reactor();
219	4017x	}));
220
221	1009x	get_resolver_service(ctx, *this);
222	1009x	get_signal_service(ctx, *this);
223	1009x	get_stream_file_service(ctx, *this);
224	1009x	get_random_access_file_service(ctx, *this);
225
226	1009x	completed_ops_.push(&task_op_);
227	1024x	}
228
229	2018x	inline epoll_scheduler::~epoll_scheduler()
230		{
231	1009x	if (timer_fd_ >= 0)
232	1009x	::close(timer_fd_);
233	1009x	if (event_fd_ >= 0)
234	1009x	::close(event_fd_);
235	1009x	if (epoll_fd_ >= 0)
236	1009x	::close(epoll_fd_);
237	2018x	}
238
239		inline void
240	1009x	epoll_scheduler::shutdown()
241		{
242	1009x	shutdown_drain();
243
244	1009x	if (event_fd_ >= 0)
245	1009x	interrupt_reactor();
246	1009x	}
247
248		inline void
249	19x	epoll_scheduler::configure_reactor(
250		unsigned max_events,
251		unsigned budget_init,
252		unsigned budget_max,
253		unsigned unassisted)
254		{
255	19x	reactor_scheduler::configure_reactor(
256		max_events, budget_init, budget_max, unassisted);
257	18x	event_buffer_.resize(max_events_per_poll_);
258	18x	}
259
260		inline std::error_code
261	7099x	epoll_scheduler::register_descriptor(int fd, reactor_descriptor_state* desc) const
262		{
263	7099x	epoll_event ev{};
264	7099x	ev.events = EPOLLIN \| EPOLLOUT \| EPOLLET \| EPOLLERR \| EPOLLHUP;
265	7099x	ev.data.ptr = desc;
266
267	7099x	if (::epoll_ctl(epoll_fd_, EPOLL_CTL_ADD, fd, &ev) < 0)
268	7x	return make_err(errno);
269
270	7092x	desc->registered_events = ev.events;
271	7092x	desc->fd = fd;
272	7092x	desc->scheduler_ = this;
273	7092x	desc->mutex.set_enabled(reactor_io_locking_);
274	7092x	desc->ready_events_.store(0, std::memory_order_relaxed);
275
276	7092x	conditionally_enabled_mutex::scoped_lock lock(desc->mutex);
277	7092x	desc->impl_ref_.reset();
278	7092x	desc->read_ready = false;
279	7092x	desc->write_ready = false;
280	7092x	return {};
281	7092x	}
282
283		inline void
284	7032x	epoll_scheduler::deregister_descriptor(int fd) const
285		{
286	7032x	::epoll_ctl(epoll_fd_, EPOLL_CTL_DEL, fd, nullptr);
287	7032x	}
288
289		inline void
290	5925x	epoll_scheduler::interrupt_reactor() const
291		{
292	5925x	bool expected = false;
293	5925x	if (eventfd_armed_.compare_exchange_strong(
294		expected, true, std::memory_order_release,
295		std::memory_order_relaxed))
296		{
297	4637x	std::uint64_t val = 1;
298	4637x	if (::write(event_fd_, &val, sizeof(val)) < 0)
299		{
300		// The flag is what coalesces later interrupts into a byte
301		// already in the eventfd; a write that failed put no byte
302		// there, so leaving it armed would swallow every interrupt
303		// that follows. Disarming keeps the cost to the interrupts
304		// already in flight -- the next one arms and writes again,
305		// instead of every one after this coalescing into a byte
306		// that does not exist.
307	2x	eventfd_armed_.store(false, std::memory_order_release);
308		}
309		}
310	5925x	}
311
312		inline void
313	6666x	epoll_scheduler::update_timerfd() const
314		{
315	6666x	auto nearest = timer_svc_->nearest_expiry();
316
317	6666x	itimerspec ts{};
318	6666x	int flags = 0;
319
320	6666x	if (nearest == timer_service::time_point::max())
321		{
322		// No timers — disarm by setting to 0 (relative)
323		}
324		else
325		{
326	6484x	auto now = std::chrono::steady_clock::now();
327	6484x	if (nearest <= now)
328		{
329		// Use 1ns instead of 0 — zero disarms the timerfd
330	363x	ts.it_value.tv_nsec = 1;
331		}
332		else
333		{
334	6121x	auto nsec = std::chrono::duration_cast<std::chrono::nanoseconds>(
335	6121x	nearest - now)
336	6121x	.count();
337	6121x	ts.it_value.tv_sec = nsec / 1000000000;
338	6121x	ts.it_value.tv_nsec = nsec % 1000000000;
339	6121x	if (ts.it_value.tv_sec == 0 && ts.it_value.tv_nsec == 0)
340	✗	ts.it_value.tv_nsec = 1;
341		}
342		}
343
344	6666x	if (::timerfd_settime(timer_fd_, flags, &ts, nullptr) < 0)
345	1x	detail::throw_system_error(make_err(errno), "timerfd_settime");
346	6665x	}
347
348		inline void
349	35780x	epoll_scheduler::run_task(
350		lock_type& lock, context_type* ctx, long timeout_us)
351		{
352		int timeout_ms;
353	35780x	if (task_interrupted_)
354	26200x	timeout_ms = 0;
355	9580x	else if (timeout_us < 0)
356	9576x	timeout_ms = -1;
357		else
358	4x	timeout_ms = static_cast<int>((timeout_us + 999) / 1000);
359
360	35780x	if (lock.owns_lock())
361	9582x	lock.unlock();
362
363	35780x	task_cleanup on_exit{this, &lock, ctx};
364
365		// Flush deferred timerfd programming before blocking
366	35780x	if (timerfd_stale_.exchange(false, std::memory_order_acquire))
367	3342x	update_timerfd();
368
369	35779x	int nfds = ::epoll_wait(
370		epoll_fd_, event_buffer_.data(),
371	35779x	static_cast<int>(event_buffer_.size()), timeout_ms);
372
373	35779x	if (nfds < 0 && errno != EINTR)
374	1x	detail::throw_system_error(make_err(errno), "epoll_wait");
375
376	35778x	bool check_timers = false;
377	35778x	ready_queue local_ops;
378
379	80571x	for (int i = 0; i < nfds; ++i)
380		{
381	44793x	if (event_buffer_[i].data.ptr == nullptr)
382		{
383		std::uint64_t val;
384		// NOLINTNEXTLINE(clang-analyzer-unix.BlockInCriticalSection)
385	3626x	[[maybe_unused]] auto r = ::read(event_fd_, &val, sizeof(val));
386	3626x	eventfd_armed_.store(false, std::memory_order_relaxed);
387	3626x	continue;
388	3626x	}
389
390	41167x	if (event_buffer_[i].data.ptr == &timer_fd_)
391		{
392		std::uint64_t expirations;
393		// NOLINTNEXTLINE(clang-analyzer-unix.BlockInCriticalSection)
394		[[maybe_unused]] auto r =
395	3324x	::read(timer_fd_, &expirations, sizeof(expirations));
396	3324x	check_timers = true;
397	3324x	continue;
398	3324x	}
399
400		auto* desc =
401	37843x	static_cast<reactor_descriptor_state*>(event_buffer_[i].data.ptr);
402	37843x	desc->add_ready_events(event_buffer_[i].events);
403
404	37843x	bool expected = false;
405	37843x	if (desc->is_enqueued_.compare_exchange_strong(
406		expected, true, std::memory_order_release,
407		std::memory_order_relaxed))
408		{
409	37843x	local_ops.push(desc);
410		}
411		}
412
413	35778x	if (check_timers)
414		{
415	3324x	timer_svc_->process_expired();
416	3324x	update_timerfd();
417		}
418
419	35778x	lock.lock();
420
421	35778x	completed_ops_.splice(local_ops);
422	35780x	}
423
424		} // namespace boost::corosio::detail
425
426		#endif // BOOST_COROSIO_HAS_EPOLL
427
428		#endif // BOOST_COROSIO_NATIVE_DETAIL_EPOLL_EPOLL_SCHEDULER_HPP
429