ustreamer/src/stream.c

/*****************************************************************************
#                                                                            #
#    uStreamer - Lightweight and fast MJPG-HTTP streamer.                    #
#                                                                            #
#    Copyright (C) 2018  Maxim Devaev <mdevaev@gmail.com>                    #
#                                                                            #
#    This program is free software: you can redistribute it and/or modify    #
#    it under the terms of the GNU General Public License as published by    #
#    the Free Software Foundation, either version 3 of the License, or       #
#    (at your option) any later version.                                     #
#                                                                            #
#    This program is distributed in the hope that it will be useful,         #
#    but WITHOUT ANY WARRANTY; without even the implied warranty of          #
#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the           #
#    GNU General Public License for more details.                            #
#                                                                            #
#    You should have received a copy of the GNU General Public License       #
#    along with this program.  If not, see <https://www.gnu.org/licenses/>.  #
#                                                                            #
*****************************************************************************/


#include <stdatomic.h>
#include <unistd.h>
#include <errno.h>
#include <time.h>
#include <assert.h>

#include <pthread.h>

#include <sys/select.h>
#include <linux/videodev2.h>

#include "tools.h"
#include "logging.h"
#include "xioctl.h"
#include "device.h"
#include "encoder.h"
#include "stream.h"


static long double _stream_get_fluency_delay(struct device_t *dev, struct workers_pool_t *pool);
static void _stream_expose_picture(struct stream_t *stream, unsigned buf_index);

static int _stream_init_loop(struct stream_t *stream, struct workers_pool_t *pool);
static int _stream_init(struct stream_t *stream, struct workers_pool_t *pool);

static void _stream_init_workers(struct stream_t *stream, struct workers_pool_t *pool);
static void *_stream_worker_thread(void *v_worker);
static void _stream_destroy_workers(struct stream_t *stream, struct workers_pool_t *pool);


struct stream_t *stream_init(struct device_t *dev, struct encoder_t *encoder) {
	struct process_t *proc;
	struct stream_t *stream;

	A_CALLOC(proc, 1);
	atomic_init(&proc->stop, false);
	atomic_init(&proc->slowdown, false);

	A_CALLOC(stream, 1);
	stream->dev = dev;
	stream->encoder = encoder;
	atomic_init(&stream->updated, false);
	A_MUTEX_INIT(&stream->mutex);
	stream->proc = proc;
	return stream;
}

void stream_destroy(struct stream_t *stream) {
	A_MUTEX_DESTROY(&stream->mutex);
	free(stream->proc);
	free(stream);
}

void stream_loop(struct stream_t *stream) {
	struct workers_pool_t pool;

	MEMSET_ZERO(pool);
	atomic_init(&pool.workers_stop, false);
	pool.encoder = stream->encoder;

	LOG_INFO("Using V4L2 device: %s", stream->dev->path);
	LOG_INFO("Using desired FPS: %u", stream->dev->desired_fps);

	while (_stream_init_loop(stream, &pool) == 0) {
		struct worker_t *oldest_worker = NULL;
		struct worker_t *last_worker = NULL;
		long double grab_after = 0;
		unsigned fluency_passed = 0;
		unsigned captured_fps_accum = 0;
		long long captured_fps_second = 0;
		bool persistent_timeout_reported = false;

		LOG_DEBUG("Allocation memory for stream picture ...");
		A_CALLOC(stream->picture.data, stream->dev->run->max_raw_image_size);

		LOG_INFO("Capturing ...");

		while (!atomic_load(&stream->proc->stop)) {
			int free_worker_number = -1;

			SEP_DEBUG('-');

			LOG_DEBUG("Waiting for workers ...");
			A_MUTEX_LOCK(&pool.free_workers_mutex);
			A_COND_WAIT_TRUE(pool.free_workers, &pool.free_workers_cond, &pool.free_workers_mutex);
			A_MUTEX_UNLOCK(&pool.free_workers_mutex);

			if (oldest_worker && !atomic_load(&oldest_worker->has_job) && oldest_worker->buf_index >= 0) {
				if (oldest_worker->job_failed) {
					break;
				}

				_stream_expose_picture(stream, oldest_worker->buf_index);

				free_worker_number = oldest_worker->number;
				oldest_worker = oldest_worker->order_next;

				LOG_PERF("##### Raw frame accepted; worker = %u", free_worker_number);
			} else {
				for (unsigned number = 0; number < stream->dev->n_workers; ++number) {
					if (
						!atomic_load(&pool.workers[number].has_job) && (
							free_worker_number == -1
							|| pool.workers[free_worker_number].job_start_time < pool.workers[number].job_start_time
						)
					) {
						free_worker_number = number;
						break;
					}
				}

				assert(free_worker_number >= 0);
				assert(!atomic_load(&pool.workers[free_worker_number].has_job));

				LOG_PERF("----- Raw frame dropped;  worker = %u", free_worker_number);
			}

			if (atomic_load(&stream->proc->stop)) {
				break;
			}

			if (atomic_load(&stream->proc->slowdown)) {
				usleep(1000000);
			}

#			define INIT_FD_SET(_set) \
				fd_set _set; FD_ZERO(&_set); FD_SET(stream->dev->run->fd, &_set);

			INIT_FD_SET(read_fds);
			INIT_FD_SET(write_fds);
			INIT_FD_SET(error_fds);

#			undef INIT_FD_SET

			struct timeval timeout;
			timeout.tv_sec = stream->dev->timeout;
			timeout.tv_usec = 0;

			LOG_DEBUG("Calling select() on video device ...");
			int retval = select(stream->dev->run->fd + 1, &read_fds, &write_fds, &error_fds, &timeout);
			LOG_DEBUG("Device select() --> %d", retval);

			if (retval < 0) {
				if (errno != EINTR) {
					LOG_PERROR("Mainloop select() error");
					break;
				}

			} else if (retval == 0) {
				if (stream->dev->persistent) {
					if (!persistent_timeout_reported) {
						LOG_ERROR("Mainloop select() timeout, polling ...")
						persistent_timeout_reported = true;
					}
					continue;
				} else {
					LOG_ERROR("Mainloop select() timeout");
					break;
				}

			} else {
				persistent_timeout_reported = false;

				if (FD_ISSET(stream->dev->run->fd, &read_fds)) {
					LOG_DEBUG("Frame is ready");

					int buf_index;
					long double now = get_now_monotonic();
					long long now_second = floor_ms(now);

					if ((buf_index = device_grab_buffer(stream->dev)) < 0) {
						break;
					}
					stream->dev->run->pictures[buf_index].grab_time = now;

					// Workaround for broken, corrupted frames:
					// Under low light conditions corrupted frames may get captured.
					// The good thing is such frames are quite small compared to the regular pictures.
					// For example a VGA (640x480) webcam picture is normally >= 8kByte large,
					// corrupted frames are smaller.
					if (stream->dev->run->hw_buffers[buf_index].used < stream->dev->min_frame_size) {
						LOG_DEBUG("Dropping too small frame sized %zu bytes, assuming it as broken",
							stream->dev->run->hw_buffers[buf_index].used);
						goto pass_frame;
					}

					{
						if (now < grab_after) {
							fluency_passed += 1;
							LOG_VERBOSE("Passed %u frames for fluency: now=%.03Lf; grab_after=%.03Lf", fluency_passed, now, grab_after);
							goto pass_frame;
						}
						fluency_passed = 0;

						if (now_second != captured_fps_second) {
							stream->captured_fps = captured_fps_accum;
							captured_fps_accum = 0;
							captured_fps_second = now_second;
							LOG_PERF("Oldest worker complete, Captured-FPS = %u", stream->captured_fps);
						}
						captured_fps_accum += 1;

						long double fluency_delay = _stream_get_fluency_delay(stream->dev, &pool);

						grab_after = now + fluency_delay;
						LOG_VERBOSE("Fluency: delay=%.03Lf; grab_after=%.03Lf", fluency_delay, grab_after);
					}

#					define FREE_WORKER(_next) pool.workers[free_worker_number]._next

					LOG_DEBUG("Grabbed a new frame to buffer %u", buf_index);

					if (!oldest_worker) {
						oldest_worker = &pool.workers[free_worker_number];
						last_worker = oldest_worker;
					} else {
						if (FREE_WORKER(order_next)) {
							FREE_WORKER(order_next->order_prev) = FREE_WORKER(order_prev);
						}
						if (FREE_WORKER(order_prev)) {
							FREE_WORKER(order_prev->order_next) = FREE_WORKER(order_next);
						}
						FREE_WORKER(order_prev) = last_worker;
						last_worker->order_next = &pool.workers[free_worker_number];
						last_worker = &pool.workers[free_worker_number];
					}
					last_worker->order_next = NULL;

					A_MUTEX_LOCK(&FREE_WORKER(has_job_mutex));
					FREE_WORKER(buf_index) = buf_index;
					atomic_store(&FREE_WORKER(has_job), true);
					A_MUTEX_UNLOCK(&FREE_WORKER(has_job_mutex));
					A_COND_SIGNAL(&FREE_WORKER(has_job_cond));

#					undef FREE_WORKER

					A_MUTEX_LOCK(&pool.free_workers_mutex);
					pool.free_workers -= 1;
					A_MUTEX_UNLOCK(&pool.free_workers_mutex);

					goto next_handlers; // Поток сам освободит буфер

					pass_frame:

					if (device_release_buffer(stream->dev, buf_index) < 0) {
						break;
					}
				}

				next_handlers:

				if (FD_ISSET(stream->dev->run->fd, &write_fds)) {
					LOG_ERROR("Got unexpected writing event, seems device was disconnected");
					break;
				}

				if (FD_ISSET(stream->dev->run->fd, &error_fds)) {
					LOG_INFO("Got V4L2 event");
					if (device_consume_event(stream->dev) < 0) {
						break;
					}
				}
			}
		}

		A_MUTEX_LOCK(&stream->mutex);
		stream->picture.used = 0; // On stream offline
		free(stream->picture.data);
		stream->width = 0;
		stream->height = 0;
		atomic_store(&stream->updated, true);
		A_MUTEX_UNLOCK(&stream->mutex);
	}

	_stream_destroy_workers(stream, &pool);
	device_switch_capturing(stream->dev, false);
	device_close(stream->dev);
}

void stream_loop_break(struct stream_t *stream) {
	atomic_store(&stream->proc->stop, true);
}

void stream_switch_slowdown(struct stream_t *stream, bool slowdown) {
	atomic_store(&stream->proc->slowdown, slowdown);
}

static void _stream_expose_picture(struct stream_t *stream, unsigned buf_index) {
#	define PICTURE(_next) stream->dev->run->pictures[buf_index]._next

	A_MUTEX_LOCK(&stream->mutex);

	stream->picture.used = PICTURE(used);
	stream->picture.allocated = PICTURE(allocated);

	memcpy(stream->picture.data, PICTURE(data), stream->picture.used * sizeof(*stream->picture.data));

	stream->picture.grab_time = PICTURE(grab_time);
	stream->picture.encode_begin_time = PICTURE(encode_begin_time);
	stream->picture.encode_end_time = PICTURE(encode_end_time);

	stream->width = stream->dev->run->width;
	stream->height = stream->dev->run->height;
	atomic_store(&stream->updated, true);

	A_MUTEX_UNLOCK(&stream->mutex);

#	undef PICTURE
}

static long double _stream_get_fluency_delay(struct device_t *dev, struct workers_pool_t *pool) {
	long double sum_comp_time = 0;
	long double avg_comp_time;
	long double min_delay;
	long double soft_delay;

	for (unsigned number = 0; number < dev->n_workers; ++number) {
#		define WORKER(_next) pool->workers[number]._next

		A_MUTEX_LOCK(&WORKER(last_comp_time_mutex));
		if (WORKER(last_comp_time) > 0) {
			sum_comp_time += WORKER(last_comp_time);
		}
		A_MUTEX_UNLOCK(&WORKER(last_comp_time_mutex));

#		undef WORKER
	}
	avg_comp_time = sum_comp_time / dev->n_workers; // Среднее время работы воркеров

	min_delay = avg_comp_time / dev->n_workers; // Среднее время работы размазывается на N воркеров

	if (dev->desired_fps > 0 && min_delay > 0) {
		// Искусственное время задержки на основе желаемого FPS, если включен --desired-fps
		soft_delay = ((long double) 1) / dev->desired_fps - sum_comp_time;
		return (min_delay > soft_delay ? min_delay : soft_delay);
	}

	return min_delay;
}

static int _stream_init_loop(struct stream_t *stream, struct workers_pool_t *pool) {
	int retval = -1;

	LOG_DEBUG("%s: stream->proc->stop = %d", __FUNCTION__, atomic_load(&stream->proc->stop));
	while (!atomic_load(&stream->proc->stop)) {
		if ((retval = _stream_init(stream, pool)) < 0) {
			LOG_INFO("Sleeping %u seconds before new stream init ...", stream->dev->error_delay);
			sleep(stream->dev->error_delay);
		} else {
			break;
		}
	}
	return retval;
}

static int _stream_init(struct stream_t *stream, struct workers_pool_t *pool) {
	SEP_INFO('=');

	_stream_destroy_workers(stream, pool);
	device_switch_capturing(stream->dev, false);
	device_close(stream->dev);

	if (device_open(stream->dev) < 0) {
		goto error;
	}
	if (device_switch_capturing(stream->dev, true) < 0) {
		goto error;
	}

	encoder_prepare_live(pool->encoder, stream->dev);

	_stream_init_workers(stream, pool);

	return 0;

	error:
		device_close(stream->dev);
		return -1;
}

static void _stream_init_workers(struct stream_t *stream, struct workers_pool_t *pool) {
	LOG_INFO("Spawning %u workers ...", stream->dev->n_workers);

	atomic_store(&pool->workers_stop, false);
	A_CALLOC(pool->workers, stream->dev->n_workers);

	A_MUTEX_INIT(&pool->free_workers_mutex);
	A_COND_INIT(&pool->free_workers_cond);

	for (unsigned number = 0; number < stream->dev->n_workers; ++number) {
#		define WORKER(_next)	pool->workers[number]._next

		pool->free_workers += 1;

		A_MUTEX_INIT(&WORKER(has_job_mutex));
		atomic_init(&WORKER(has_job), false);
		A_COND_INIT(&WORKER(has_job_cond));

		WORKER(number)			= number;
		WORKER(proc_stop)		= &stream->proc->stop;
		WORKER(workers_stop)	= &pool->workers_stop;

		WORKER(free_workers_mutex)	= &pool->free_workers_mutex;
		WORKER(free_workers)		= &pool->free_workers;
		WORKER(free_workers_cond)	= &pool->free_workers_cond;

		WORKER(dev)		= stream->dev;
		WORKER(encoder)	= pool->encoder;

		A_THREAD_CREATE(&WORKER(tid), _stream_worker_thread, (void *)&(pool->workers[number]));

#		undef WORKER
	}
}

static void *_stream_worker_thread(void *v_worker) {
	struct worker_t *worker = (struct worker_t *)v_worker;

	LOG_DEBUG("Hello! I am a worker #%u ^_^", worker->number);

	while (!atomic_load(worker->proc_stop) && !atomic_load(worker->workers_stop)) {
		LOG_DEBUG("Worker %u waiting for a new job ...", worker->number);
		A_MUTEX_LOCK(&worker->has_job_mutex);
		A_COND_WAIT_TRUE(atomic_load(&worker->has_job), &worker->has_job_cond, &worker->has_job_mutex);
		A_MUTEX_UNLOCK(&worker->has_job_mutex);

		if (!atomic_load(worker->workers_stop)) {
#			define PICTURE(_next) worker->dev->run->pictures[worker->buf_index]._next

			LOG_DEBUG("Worker %u compressing JPEG from buffer %u ...", worker->number, worker->buf_index);

			PICTURE(encode_begin_time) = get_now_monotonic();
			if (encoder_compress_buffer(worker->encoder, worker->dev, worker->number, worker->buf_index) < 0) {
				worker->job_failed = true;
			}
			PICTURE(encode_end_time) = get_now_monotonic();

			if (device_release_buffer(worker->dev, worker->buf_index) == 0) {
				worker->job_start_time = PICTURE(encode_begin_time);
				atomic_store(&worker->has_job, false);

				long double last_comp_time = PICTURE(encode_end_time) - worker->job_start_time;

				A_MUTEX_LOCK(&worker->last_comp_time_mutex);
				worker->last_comp_time = last_comp_time;
				A_MUTEX_UNLOCK(&worker->last_comp_time_mutex);

				LOG_VERBOSE("Compressed JPEG size=%zu; time=%0.3Lf; worker=%u; buffer=%u",
					PICTURE(used), last_comp_time, worker->number, worker->buf_index);
			} else {
				worker->job_failed = true;
				atomic_store(&worker->has_job, false);
			}

#			undef PICTURE
		}

		A_MUTEX_LOCK(worker->free_workers_mutex);
		*worker->free_workers += 1;
		A_MUTEX_UNLOCK(worker->free_workers_mutex);
		A_COND_SIGNAL(worker->free_workers_cond);
	}

	LOG_DEBUG("Bye-bye (worker %u)", worker->number);
	return NULL;
}

static void _stream_destroy_workers(struct stream_t *stream, struct workers_pool_t *pool) {
	if (pool->workers) {
		LOG_INFO("Destroying workers ...");

		atomic_store(&pool->workers_stop, true);
		for (unsigned number = 0; number < stream->dev->n_workers; ++number) {
#			define WORKER(_next) pool->workers[number]._next

			A_MUTEX_LOCK(&WORKER(has_job_mutex));
			atomic_store(&WORKER(has_job), true); // Final job: die
			A_MUTEX_UNLOCK(&WORKER(has_job_mutex));
			A_COND_SIGNAL(&WORKER(has_job_cond));

			A_THREAD_JOIN(WORKER(tid));
			A_MUTEX_DESTROY(&WORKER(has_job_mutex));
			A_COND_DESTROY(&WORKER(has_job_cond));

#			undef WORKER
		}

		A_MUTEX_DESTROY(&pool->free_workers_mutex);
		A_COND_DESTROY(&pool->free_workers_cond);

		free(pool->workers);
	}
	pool->free_workers = 0;
	pool->workers = NULL;
}