/*
- * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (c) 2004 Roman Shaposhnik
+ * Copyright (c) 2008 Alexander Strange (astrange@ithinksw.com)
*
- * This library is free software; you can redistribute it and/or
+ * Many thanks to Steven M. Schultz for providing clever ideas and
+ * to Michael Niedermayer <michaelni@gmx.at> for writing initial
+ * implementation.
+ *
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
+ * version 2.1 of the License, or (at your option) any later version.
*
- * This library is distributed in the hope that it will be useful,
+ * Libav 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
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
- *
+ * License along with Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
-#include <semaphore.h>
-#include <pthread.h>
-//#define DEBUG
+/**
+ * @file
+ * Multithreading support functions
+ * @see doc/multithreading.txt
+ */
+
+#include "config.h"
+
+#if HAVE_SCHED_GETAFFINITY
+#define _GNU_SOURCE
+#include <sched.h>
+#endif
+#if HAVE_GETPROCESSAFFINITYMASK
+#include <windows.h>
+#endif
+#if HAVE_SYSCTL
+#if HAVE_SYS_PARAM_H
+#include <sys/param.h>
+#endif
+#include <sys/types.h>
+#include <sys/sysctl.h>
+#endif
+#if HAVE_SYSCONF
+#include <unistd.h>
+#endif
#include "avcodec.h"
-#include "common.h"
-
-typedef struct JobContext{
- sem_t available_sem;
- int assigned;
- int (*func)(AVCodecContext *c, void *arg);
- void *arg;
- int ret;
-}JobContext;
-
-typedef struct WorkerContext{
- AVCodecContext *avctx;
- pthread_t thread;
- int start_index;
- sem_t work_sem;
- sem_t done_sem;
-}WorkerContext;
-
-typedef struct ThreadContext{
- WorkerContext *worker;
- JobContext *job;
+#include "internal.h"
+#include "thread.h"
+
+#if HAVE_PTHREADS
+#include <pthread.h>
+#elif HAVE_W32THREADS
+#include "w32pthreads.h"
+#endif
+
+typedef int (action_func)(AVCodecContext *c, void *arg);
+typedef int (action_func2)(AVCodecContext *c, void *arg, int jobnr, int threadnr);
+
+typedef struct ThreadContext {
+ pthread_t *workers;
+ action_func *func;
+ action_func2 *func2;
+ void *args;
+ int *rets;
+ int rets_count;
int job_count;
- int allocated_job_count;
-}ThreadContext;
+ int job_size;
-static void * thread_func(void *v){
- WorkerContext *w= v;
- ThreadContext *c= w->avctx->thread_opaque;
- int i;
+ pthread_cond_t last_job_cond;
+ pthread_cond_t current_job_cond;
+ pthread_mutex_t current_job_lock;
+ int current_job;
+ int done;
+} ThreadContext;
+
+/// Max number of frame buffers that can be allocated when using frame threads.
+#define MAX_BUFFERS (32+1)
+
+/**
+ * Context used by codec threads and stored in their AVCodecContext thread_opaque.
+ */
+typedef struct PerThreadContext {
+ struct FrameThreadContext *parent;
+
+ pthread_t thread;
+ int thread_init;
+ pthread_cond_t input_cond; ///< Used to wait for a new packet from the main thread.
+ pthread_cond_t progress_cond; ///< Used by child threads to wait for progress to change.
+ pthread_cond_t output_cond; ///< Used by the main thread to wait for frames to finish.
+
+ pthread_mutex_t mutex; ///< Mutex used to protect the contents of the PerThreadContext.
+ pthread_mutex_t progress_mutex; ///< Mutex used to protect frame progress values and progress_cond.
+
+ AVCodecContext *avctx; ///< Context used to decode packets passed to this thread.
+
+ AVPacket avpkt; ///< Input packet (for decoding) or output (for encoding).
+ int allocated_buf_size; ///< Size allocated for avpkt.data
+
+ AVFrame frame; ///< Output frame (for decoding) or input (for encoding).
+ int got_frame; ///< The output of got_picture_ptr from the last avcodec_decode_video() call.
+ int result; ///< The result of the last codec decode/encode() call.
+
+ enum {
+ STATE_INPUT_READY, ///< Set when the thread is awaiting a packet.
+ STATE_SETTING_UP, ///< Set before the codec has called ff_thread_finish_setup().
+ STATE_GET_BUFFER, /**<
+ * Set when the codec calls get_buffer().
+ * State is returned to STATE_SETTING_UP afterwards.
+ */
+ STATE_SETUP_FINISHED ///< Set after the codec has called ff_thread_finish_setup().
+ } state;
+
+ /**
+ * Array of frames passed to ff_thread_release_buffer().
+ * Frames are released after all threads referencing them are finished.
+ */
+ AVFrame released_buffers[MAX_BUFFERS];
+ int num_released_buffers;
+
+ /**
+ * Array of progress values used by ff_thread_get_buffer().
+ */
+ int progress[MAX_BUFFERS][2];
+ uint8_t progress_used[MAX_BUFFERS];
+
+ AVFrame *requested_frame; ///< AVFrame the codec passed to get_buffer()
+} PerThreadContext;
+
+/**
+ * Context stored in the client AVCodecContext thread_opaque.
+ */
+typedef struct FrameThreadContext {
+ PerThreadContext *threads; ///< The contexts for each thread.
+ PerThreadContext *prev_thread; ///< The last thread submit_packet() was called on.
+
+ pthread_mutex_t buffer_mutex; ///< Mutex used to protect get/release_buffer().
+
+ int next_decoding; ///< The next context to submit a packet to.
+ int next_finished; ///< The next context to return output from.
+
+ int delaying; /**<
+ * Set for the first N packets, where N is the number of threads.
+ * While it is set, ff_thread_en/decode_frame won't return any results.
+ */
+
+ int die; ///< Set when threads should exit.
+} FrameThreadContext;
+
+
+/* H264 slice threading seems to be buggy with more than 16 threads,
+ * limit the number of threads to 16 for automatic detection */
+#define MAX_AUTO_THREADS 16
+
+static int get_logical_cpus(AVCodecContext *avctx)
+{
+ int ret, nb_cpus = 1;
+#if HAVE_SCHED_GETAFFINITY && defined(CPU_COUNT)
+ cpu_set_t cpuset;
- for(;;){
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X enter wait\n", (int)v);
- sem_wait(&w->work_sem);
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X after wait\n", (int)v);
- if(c->job_count == 0)
- break;
-
- for(i=0; i<c->job_count; i++){
- int index= (i + w->start_index) % c->job_count;
- JobContext *j= &c->job[index];
-
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X first check of %d\n", (int)v, index);
- if(j->assigned) continue; //unsynced check, if != 0 it is already given to another worker, it never becomes available before the next execute() call so this should be safe
-
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X second check of %d\n", (int)v, index);
- if(sem_trywait(&j->available_sem) == 0){
- j->assigned=1;
- j->ret= j->func(w->avctx, j->arg);
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X done %d\n", (int)v, index);
+ CPU_ZERO(&cpuset);
+
+ ret = sched_getaffinity(0, sizeof(cpuset), &cpuset);
+ if (!ret) {
+ nb_cpus = CPU_COUNT(&cpuset);
+ }
+#elif HAVE_GETPROCESSAFFINITYMASK
+ DWORD_PTR proc_aff, sys_aff;
+ ret = GetProcessAffinityMask(GetCurrentProcess(), &proc_aff, &sys_aff);
+ if (ret)
+ nb_cpus = av_popcount64(proc_aff);
+#elif HAVE_SYSCTL && defined(HW_NCPU)
+ int mib[2] = { CTL_HW, HW_NCPU };
+ size_t len = sizeof(nb_cpus);
+
+ ret = sysctl(mib, 2, &nb_cpus, &len, NULL, 0);
+ if (ret == -1)
+ nb_cpus = 0;
+#elif HAVE_SYSCONF && defined(_SC_NPROC_ONLN)
+ nb_cpus = sysconf(_SC_NPROC_ONLN);
+#elif HAVE_SYSCONF && defined(_SC_NPROCESSORS_ONLN)
+ nb_cpus = sysconf(_SC_NPROCESSORS_ONLN);
+#endif
+ av_log(avctx, AV_LOG_DEBUG, "detected %d logical cores\n", nb_cpus);
+ return nb_cpus;
+}
+
+
+static void* attribute_align_arg worker(void *v)
+{
+ AVCodecContext *avctx = v;
+ ThreadContext *c = avctx->thread_opaque;
+ int our_job = c->job_count;
+ int thread_count = avctx->thread_count;
+ int self_id;
+
+ pthread_mutex_lock(&c->current_job_lock);
+ self_id = c->current_job++;
+ for (;;){
+ while (our_job >= c->job_count) {
+ if (c->current_job == thread_count + c->job_count)
+ pthread_cond_signal(&c->last_job_cond);
+
+ pthread_cond_wait(&c->current_job_cond, &c->current_job_lock);
+ our_job = self_id;
+
+ if (c->done) {
+ pthread_mutex_unlock(&c->current_job_lock);
+ return NULL;
}
}
-//av_log(w->avctx, AV_LOG_DEBUG, "thread_func %X complete\n", (int)v);
- sem_post(&w->done_sem);
+ pthread_mutex_unlock(&c->current_job_lock);
+
+ c->rets[our_job%c->rets_count] = c->func ? c->func(avctx, (char*)c->args + our_job*c->job_size):
+ c->func2(avctx, c->args, our_job, self_id);
+
+ pthread_mutex_lock(&c->current_job_lock);
+ our_job = c->current_job++;
+ }
+}
+
+static av_always_inline void avcodec_thread_park_workers(ThreadContext *c, int thread_count)
+{
+ pthread_cond_wait(&c->last_job_cond, &c->current_job_lock);
+ pthread_mutex_unlock(&c->current_job_lock);
+}
+
+static void thread_free(AVCodecContext *avctx)
+{
+ ThreadContext *c = avctx->thread_opaque;
+ int i;
+
+ pthread_mutex_lock(&c->current_job_lock);
+ c->done = 1;
+ pthread_cond_broadcast(&c->current_job_cond);
+ pthread_mutex_unlock(&c->current_job_lock);
+
+ for (i=0; i<avctx->thread_count; i++)
+ pthread_join(c->workers[i], NULL);
+
+ pthread_mutex_destroy(&c->current_job_lock);
+ pthread_cond_destroy(&c->current_job_cond);
+ pthread_cond_destroy(&c->last_job_cond);
+ av_free(c->workers);
+ av_freep(&avctx->thread_opaque);
+}
+
+static int avcodec_thread_execute(AVCodecContext *avctx, action_func* func, void *arg, int *ret, int job_count, int job_size)
+{
+ ThreadContext *c= avctx->thread_opaque;
+ int dummy_ret;
+
+ if (!(avctx->active_thread_type&FF_THREAD_SLICE) || avctx->thread_count <= 1)
+ return avcodec_default_execute(avctx, func, arg, ret, job_count, job_size);
+
+ if (job_count <= 0)
+ return 0;
+
+ pthread_mutex_lock(&c->current_job_lock);
+
+ c->current_job = avctx->thread_count;
+ c->job_count = job_count;
+ c->job_size = job_size;
+ c->args = arg;
+ c->func = func;
+ if (ret) {
+ c->rets = ret;
+ c->rets_count = job_count;
+ } else {
+ c->rets = &dummy_ret;
+ c->rets_count = 1;
+ }
+ pthread_cond_broadcast(&c->current_job_cond);
+
+ avcodec_thread_park_workers(c, avctx->thread_count);
+
+ return 0;
+}
+
+static int avcodec_thread_execute2(AVCodecContext *avctx, action_func2* func2, void *arg, int *ret, int job_count)
+{
+ ThreadContext *c= avctx->thread_opaque;
+ c->func2 = func2;
+ return avcodec_thread_execute(avctx, NULL, arg, ret, job_count, 0);
+}
+
+static int thread_init(AVCodecContext *avctx)
+{
+ int i;
+ ThreadContext *c;
+ int thread_count = avctx->thread_count;
+
+ if (!thread_count) {
+ int nb_cpus = get_logical_cpus(avctx);
+ // use number of cores + 1 as thread count if there is more than one
+ if (nb_cpus > 1)
+ thread_count = avctx->thread_count = FFMIN(nb_cpus + 1, MAX_AUTO_THREADS);
+ else
+ thread_count = avctx->thread_count = 1;
+ }
+
+ if (thread_count <= 1) {
+ avctx->active_thread_type = 0;
+ return 0;
+ }
+
+ c = av_mallocz(sizeof(ThreadContext));
+ if (!c)
+ return -1;
+
+ c->workers = av_mallocz(sizeof(pthread_t)*thread_count);
+ if (!c->workers) {
+ av_free(c);
+ return -1;
+ }
+
+ avctx->thread_opaque = c;
+ c->current_job = 0;
+ c->job_count = 0;
+ c->job_size = 0;
+ c->done = 0;
+ pthread_cond_init(&c->current_job_cond, NULL);
+ pthread_cond_init(&c->last_job_cond, NULL);
+ pthread_mutex_init(&c->current_job_lock, NULL);
+ pthread_mutex_lock(&c->current_job_lock);
+ for (i=0; i<thread_count; i++) {
+ if(pthread_create(&c->workers[i], NULL, worker, avctx)) {
+ avctx->thread_count = i;
+ pthread_mutex_unlock(&c->current_job_lock);
+ ff_thread_free(avctx);
+ return -1;
+ }
+ }
+
+ avcodec_thread_park_workers(c, thread_count);
+
+ avctx->execute = avcodec_thread_execute;
+ avctx->execute2 = avcodec_thread_execute2;
+ return 0;
+}
+
+/**
+ * Codec worker thread.
+ *
+ * Automatically calls ff_thread_finish_setup() if the codec does
+ * not provide an update_thread_context method, or if the codec returns
+ * before calling it.
+ */
+static attribute_align_arg void *frame_worker_thread(void *arg)
+{
+ PerThreadContext *p = arg;
+ FrameThreadContext *fctx = p->parent;
+ AVCodecContext *avctx = p->avctx;
+ AVCodec *codec = avctx->codec;
+
+ while (1) {
+ if (p->state == STATE_INPUT_READY && !fctx->die) {
+ pthread_mutex_lock(&p->mutex);
+ while (p->state == STATE_INPUT_READY && !fctx->die)
+ pthread_cond_wait(&p->input_cond, &p->mutex);
+ pthread_mutex_unlock(&p->mutex);
+ }
+
+ if (fctx->die) break;
+
+ if (!codec->update_thread_context && avctx->thread_safe_callbacks)
+ ff_thread_finish_setup(avctx);
+
+ pthread_mutex_lock(&p->mutex);
+ avcodec_get_frame_defaults(&p->frame);
+ p->got_frame = 0;
+ p->result = codec->decode(avctx, &p->frame, &p->got_frame, &p->avpkt);
+
+ if (p->state == STATE_SETTING_UP) ff_thread_finish_setup(avctx);
+
+ p->state = STATE_INPUT_READY;
+
+ pthread_mutex_lock(&p->progress_mutex);
+ pthread_cond_signal(&p->output_cond);
+ pthread_mutex_unlock(&p->progress_mutex);
+
+ pthread_mutex_unlock(&p->mutex);
}
-
+
return NULL;
}
/**
- * free what has been allocated by avcodec_thread_init().
- * must be called after decoding has finished, especially dont call while avcodec_thread_execute() is running
+ * Update the next thread's AVCodecContext with values from the reference thread's context.
+ *
+ * @param dst The destination context.
+ * @param src The source context.
+ * @param for_user 0 if the destination is a codec thread, 1 if the destination is the user's thread
*/
-void avcodec_thread_free(AVCodecContext *s){
- ThreadContext *c= s->thread_opaque;
- int i, val;
-
- for(i=0; i<c->allocated_job_count; i++){
- sem_getvalue(&c->job[i].available_sem, &val); assert(val == 0);
- sem_destroy(&c->job[i].available_sem);
+static int update_context_from_thread(AVCodecContext *dst, AVCodecContext *src, int for_user)
+{
+ int err = 0;
+
+ if (dst != src) {
+ dst->time_base = src->time_base;
+ dst->width = src->width;
+ dst->height = src->height;
+ dst->pix_fmt = src->pix_fmt;
+
+ dst->coded_width = src->coded_width;
+ dst->coded_height = src->coded_height;
+
+ dst->has_b_frames = src->has_b_frames;
+ dst->idct_algo = src->idct_algo;
+
+ dst->bits_per_coded_sample = src->bits_per_coded_sample;
+ dst->sample_aspect_ratio = src->sample_aspect_ratio;
+ dst->dtg_active_format = src->dtg_active_format;
+
+ dst->profile = src->profile;
+ dst->level = src->level;
+
+ dst->bits_per_raw_sample = src->bits_per_raw_sample;
+ dst->ticks_per_frame = src->ticks_per_frame;
+ dst->color_primaries = src->color_primaries;
+
+ dst->color_trc = src->color_trc;
+ dst->colorspace = src->colorspace;
+ dst->color_range = src->color_range;
+ dst->chroma_sample_location = src->chroma_sample_location;
+ }
+
+ if (for_user) {
+ dst->coded_frame = src->coded_frame;
+ } else {
+ if (dst->codec->update_thread_context)
+ err = dst->codec->update_thread_context(dst, src);
}
- c->job_count= 0;
- for(i=0; i<s->thread_count; i++){
- sem_getvalue(&c->worker[i].work_sem, &val); assert(val == 0);
- sem_getvalue(&c->worker[i].done_sem, &val); assert(val == 0);
+ return err;
+}
+
+/**
+ * Update the next thread's AVCodecContext with values set by the user.
+ *
+ * @param dst The destination context.
+ * @param src The source context.
+ * @return 0 on success, negative error code on failure
+ */
+static int update_context_from_user(AVCodecContext *dst, AVCodecContext *src)
+{
+#define copy_fields(s, e) memcpy(&dst->s, &src->s, (char*)&dst->e - (char*)&dst->s);
+ dst->flags = src->flags;
+
+ dst->draw_horiz_band= src->draw_horiz_band;
+ dst->get_buffer = src->get_buffer;
+ dst->release_buffer = src->release_buffer;
- sem_post(&c->worker[i].work_sem);
- pthread_join(c->worker[i].thread, NULL);
- sem_destroy(&c->worker[i].work_sem);
- sem_destroy(&c->worker[i].done_sem);
+ dst->opaque = src->opaque;
+ dst->debug = src->debug;
+ dst->debug_mv = src->debug_mv;
+
+ dst->slice_flags = src->slice_flags;
+ dst->flags2 = src->flags2;
+
+ copy_fields(skip_loop_filter, subtitle_header);
+
+ dst->frame_number = src->frame_number;
+ dst->reordered_opaque = src->reordered_opaque;
+
+ if (src->slice_count && src->slice_offset) {
+ if (dst->slice_count < src->slice_count) {
+ int *tmp = av_realloc(dst->slice_offset, src->slice_count *
+ sizeof(*dst->slice_offset));
+ if (!tmp) {
+ av_free(dst->slice_offset);
+ return AVERROR(ENOMEM);
+ }
+ dst->slice_offset = tmp;
+ }
+ memcpy(dst->slice_offset, src->slice_offset,
+ src->slice_count * sizeof(*dst->slice_offset));
}
+ dst->slice_count = src->slice_count;
+ return 0;
+#undef copy_fields
+}
+
+static void free_progress(AVFrame *f)
+{
+ PerThreadContext *p = f->owner->thread_opaque;
+ int *progress = f->thread_opaque;
- av_freep(&c->job);
- av_freep(&c->worker);
- av_freep(&s->thread_opaque);
+ p->progress_used[(progress - p->progress[0]) / 2] = 0;
}
-int avcodec_thread_execute(AVCodecContext *s, int (*func)(AVCodecContext *c2, void *arg2),void **arg, int *ret, int job_count){
- ThreadContext *c= s->thread_opaque;
- int i, val;
-
- assert(s == c->avctx);
- if(job_count > c->allocated_job_count){
- c->job= av_realloc(c->job, job_count*sizeof(JobContext));
+/// Releases the buffers that this decoding thread was the last user of.
+static void release_delayed_buffers(PerThreadContext *p)
+{
+ FrameThreadContext *fctx = p->parent;
- for(i=c->allocated_job_count; i<job_count; i++){
- memset(&c->job[i], 0, sizeof(JobContext));
- c->allocated_job_count++;
+ while (p->num_released_buffers > 0) {
+ AVFrame *f;
- if(sem_init(&c->job[i].available_sem, 0, 0))
- return -1;
+ pthread_mutex_lock(&fctx->buffer_mutex);
+ f = &p->released_buffers[--p->num_released_buffers];
+ free_progress(f);
+ f->thread_opaque = NULL;
+
+ f->owner->release_buffer(f->owner, f);
+ pthread_mutex_unlock(&fctx->buffer_mutex);
+ }
+}
+
+static int submit_packet(PerThreadContext *p, AVPacket *avpkt)
+{
+ FrameThreadContext *fctx = p->parent;
+ PerThreadContext *prev_thread = fctx->prev_thread;
+ AVCodec *codec = p->avctx->codec;
+ uint8_t *buf = p->avpkt.data;
+
+ if (!avpkt->size && !(codec->capabilities & CODEC_CAP_DELAY)) return 0;
+
+ pthread_mutex_lock(&p->mutex);
+
+ release_delayed_buffers(p);
+
+ if (prev_thread) {
+ int err;
+ if (prev_thread->state == STATE_SETTING_UP) {
+ pthread_mutex_lock(&prev_thread->progress_mutex);
+ while (prev_thread->state == STATE_SETTING_UP)
+ pthread_cond_wait(&prev_thread->progress_cond, &prev_thread->progress_mutex);
+ pthread_mutex_unlock(&prev_thread->progress_mutex);
+ }
+
+ err = update_context_from_thread(p->avctx, prev_thread->avctx, 0);
+ if (err) {
+ pthread_mutex_unlock(&p->mutex);
+ return err;
}
}
- c->job_count= job_count;
-
- /* note, we can be certain that this is not called with the same AVCodecContext by different threads at the same time */
- for(i=0; i<job_count; i++){
- sem_getvalue(&c->job[i].available_sem, &val); assert(val == 0);
-
- c->job[i].arg= arg[i];
- c->job[i].func= func;
- c->job[i].ret= 12345;
- c->job[i].assigned= 0;
- sem_post(&c->job[i].available_sem);
+ av_fast_malloc(&buf, &p->allocated_buf_size, avpkt->size + FF_INPUT_BUFFER_PADDING_SIZE);
+ p->avpkt = *avpkt;
+ p->avpkt.data = buf;
+ memcpy(buf, avpkt->data, avpkt->size);
+ memset(buf + avpkt->size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
+
+ p->state = STATE_SETTING_UP;
+ pthread_cond_signal(&p->input_cond);
+ pthread_mutex_unlock(&p->mutex);
+
+ /*
+ * If the client doesn't have a thread-safe get_buffer(),
+ * then decoding threads call back to the main thread,
+ * and it calls back to the client here.
+ */
+
+ if (!p->avctx->thread_safe_callbacks &&
+ p->avctx->get_buffer != avcodec_default_get_buffer) {
+ while (p->state != STATE_SETUP_FINISHED && p->state != STATE_INPUT_READY) {
+ pthread_mutex_lock(&p->progress_mutex);
+ while (p->state == STATE_SETTING_UP)
+ pthread_cond_wait(&p->progress_cond, &p->progress_mutex);
+
+ if (p->state == STATE_GET_BUFFER) {
+ p->result = p->avctx->get_buffer(p->avctx, p->requested_frame);
+ p->state = STATE_SETTING_UP;
+ pthread_cond_signal(&p->progress_cond);
+ }
+ pthread_mutex_unlock(&p->progress_mutex);
+ }
}
- for(i=0; i<s->thread_count && i<job_count; i++){
- sem_getvalue(&c->worker[i].work_sem, &val); assert(val == 0);
- sem_getvalue(&c->worker[i].done_sem, &val); assert(val == 0);
+ fctx->prev_thread = p;
+ fctx->next_decoding++;
- c->worker[i].start_index= (i + job_count/2)/job_count;
-//av_log(s, AV_LOG_DEBUG, "start worker %d\n", i);
- sem_post(&c->worker[i].work_sem);
+ return 0;
+}
+
+int ff_thread_decode_frame(AVCodecContext *avctx,
+ AVFrame *picture, int *got_picture_ptr,
+ AVPacket *avpkt)
+{
+ FrameThreadContext *fctx = avctx->thread_opaque;
+ int finished = fctx->next_finished;
+ PerThreadContext *p;
+ int err;
+
+ /*
+ * Submit a packet to the next decoding thread.
+ */
+
+ p = &fctx->threads[fctx->next_decoding];
+ err = update_context_from_user(p->avctx, avctx);
+ if (err) return err;
+ err = submit_packet(p, avpkt);
+ if (err) return err;
+
+ /*
+ * If we're still receiving the initial packets, don't return a frame.
+ */
+
+ if (fctx->delaying) {
+ if (fctx->next_decoding >= (avctx->thread_count-1)) fctx->delaying = 0;
+
+ *got_picture_ptr=0;
+ if (avpkt->size)
+ return avpkt->size;
}
- for(i=0; i<s->thread_count && i<job_count; i++){
-//av_log(s, AV_LOG_DEBUG, "wait for worker %d\n", i);
- sem_wait(&c->worker[i].done_sem);
+ /*
+ * Return the next available frame from the oldest thread.
+ * If we're at the end of the stream, then we have to skip threads that
+ * didn't output a frame, because we don't want to accidentally signal
+ * EOF (avpkt->size == 0 && *got_picture_ptr == 0).
+ */
+
+ do {
+ p = &fctx->threads[finished++];
+
+ if (p->state != STATE_INPUT_READY) {
+ pthread_mutex_lock(&p->progress_mutex);
+ while (p->state != STATE_INPUT_READY)
+ pthread_cond_wait(&p->output_cond, &p->progress_mutex);
+ pthread_mutex_unlock(&p->progress_mutex);
+ }
+
+ *picture = p->frame;
+ *got_picture_ptr = p->got_frame;
+ picture->pkt_dts = p->avpkt.dts;
+
+ /*
+ * A later call with avkpt->size == 0 may loop over all threads,
+ * including this one, searching for a frame to return before being
+ * stopped by the "finished != fctx->next_finished" condition.
+ * Make sure we don't mistakenly return the same frame again.
+ */
+ p->got_frame = 0;
+
+ if (finished >= avctx->thread_count) finished = 0;
+ } while (!avpkt->size && !*got_picture_ptr && finished != fctx->next_finished);
+
+ update_context_from_thread(avctx, p->avctx, 1);
+
+ if (fctx->next_decoding >= avctx->thread_count) fctx->next_decoding = 0;
+
+ fctx->next_finished = finished;
+
+ /* return the size of the consumed packet if no error occurred */
+ return (p->result >= 0) ? avpkt->size : p->result;
+}
+
+void ff_thread_report_progress(AVFrame *f, int n, int field)
+{
+ PerThreadContext *p;
+ int *progress = f->thread_opaque;
+
+ if (!progress || progress[field] >= n) return;
+
+ p = f->owner->thread_opaque;
+
+ if (f->owner->debug&FF_DEBUG_THREADS)
+ av_log(f->owner, AV_LOG_DEBUG, "%p finished %d field %d\n", progress, n, field);
+
+ pthread_mutex_lock(&p->progress_mutex);
+ progress[field] = n;
+ pthread_cond_broadcast(&p->progress_cond);
+ pthread_mutex_unlock(&p->progress_mutex);
+}
+
+void ff_thread_await_progress(AVFrame *f, int n, int field)
+{
+ PerThreadContext *p;
+ int *progress = f->thread_opaque;
+
+ if (!progress || progress[field] >= n) return;
+
+ p = f->owner->thread_opaque;
+
+ if (f->owner->debug&FF_DEBUG_THREADS)
+ av_log(f->owner, AV_LOG_DEBUG, "thread awaiting %d field %d from %p\n", n, field, progress);
+
+ pthread_mutex_lock(&p->progress_mutex);
+ while (progress[field] < n)
+ pthread_cond_wait(&p->progress_cond, &p->progress_mutex);
+ pthread_mutex_unlock(&p->progress_mutex);
+}
+
+void ff_thread_finish_setup(AVCodecContext *avctx) {
+ PerThreadContext *p = avctx->thread_opaque;
+
+ if (!(avctx->active_thread_type&FF_THREAD_FRAME)) return;
+
+ pthread_mutex_lock(&p->progress_mutex);
+ p->state = STATE_SETUP_FINISHED;
+ pthread_cond_broadcast(&p->progress_cond);
+ pthread_mutex_unlock(&p->progress_mutex);
+}
+
+/// Waits for all threads to finish.
+static void park_frame_worker_threads(FrameThreadContext *fctx, int thread_count)
+{
+ int i;
- sem_getvalue(&c->worker[i].work_sem, &val); assert(val == 0);
- sem_getvalue(&c->worker[i].done_sem, &val); assert(val == 0);
+ for (i = 0; i < thread_count; i++) {
+ PerThreadContext *p = &fctx->threads[i];
+
+ if (p->state != STATE_INPUT_READY) {
+ pthread_mutex_lock(&p->progress_mutex);
+ while (p->state != STATE_INPUT_READY)
+ pthread_cond_wait(&p->output_cond, &p->progress_mutex);
+ pthread_mutex_unlock(&p->progress_mutex);
+ }
}
+}
+
+static void frame_thread_free(AVCodecContext *avctx, int thread_count)
+{
+ FrameThreadContext *fctx = avctx->thread_opaque;
+ AVCodec *codec = avctx->codec;
+ int i;
+
+ park_frame_worker_threads(fctx, thread_count);
+
+ if (fctx->prev_thread && fctx->prev_thread != fctx->threads)
+ update_context_from_thread(fctx->threads->avctx, fctx->prev_thread->avctx, 0);
+
+ fctx->die = 1;
+
+ for (i = 0; i < thread_count; i++) {
+ PerThreadContext *p = &fctx->threads[i];
- for(i=0; i<job_count; i++){
- sem_getvalue(&c->job[i].available_sem, &val); assert(val == 0);
-
- c->job[i].func= NULL;
- if(ret) ret[i]= c->job[i].ret;
+ pthread_mutex_lock(&p->mutex);
+ pthread_cond_signal(&p->input_cond);
+ pthread_mutex_unlock(&p->mutex);
+
+ if (p->thread_init)
+ pthread_join(p->thread, NULL);
+
+ if (codec->close)
+ codec->close(p->avctx);
+
+ avctx->codec = NULL;
+
+ release_delayed_buffers(p);
+ }
+
+ for (i = 0; i < thread_count; i++) {
+ PerThreadContext *p = &fctx->threads[i];
+
+ avcodec_default_free_buffers(p->avctx);
+
+ pthread_mutex_destroy(&p->mutex);
+ pthread_mutex_destroy(&p->progress_mutex);
+ pthread_cond_destroy(&p->input_cond);
+ pthread_cond_destroy(&p->progress_cond);
+ pthread_cond_destroy(&p->output_cond);
+ av_freep(&p->avpkt.data);
+
+ if (i) {
+ av_freep(&p->avctx->priv_data);
+ av_freep(&p->avctx->internal);
+ av_freep(&p->avctx->slice_offset);
+ }
+
+ av_freep(&p->avctx);
+ }
+
+ av_freep(&fctx->threads);
+ pthread_mutex_destroy(&fctx->buffer_mutex);
+ av_freep(&avctx->thread_opaque);
+}
+
+static int frame_thread_init(AVCodecContext *avctx)
+{
+ int thread_count = avctx->thread_count;
+ AVCodec *codec = avctx->codec;
+ AVCodecContext *src = avctx;
+ FrameThreadContext *fctx;
+ int i, err = 0;
+
+ if (!thread_count) {
+ int nb_cpus = get_logical_cpus(avctx);
+ // use number of cores + 1 as thread count if there is more than one
+ if (nb_cpus > 1)
+ thread_count = avctx->thread_count = FFMIN(nb_cpus + 1, MAX_AUTO_THREADS);
+ else
+ thread_count = avctx->thread_count = 1;
+ }
+
+ if (thread_count <= 1) {
+ avctx->active_thread_type = 0;
+ return 0;
+ }
+
+ avctx->thread_opaque = fctx = av_mallocz(sizeof(FrameThreadContext));
+
+ fctx->threads = av_mallocz(sizeof(PerThreadContext) * thread_count);
+ pthread_mutex_init(&fctx->buffer_mutex, NULL);
+ fctx->delaying = 1;
+
+ for (i = 0; i < thread_count; i++) {
+ AVCodecContext *copy = av_malloc(sizeof(AVCodecContext));
+ PerThreadContext *p = &fctx->threads[i];
+
+ pthread_mutex_init(&p->mutex, NULL);
+ pthread_mutex_init(&p->progress_mutex, NULL);
+ pthread_cond_init(&p->input_cond, NULL);
+ pthread_cond_init(&p->progress_cond, NULL);
+ pthread_cond_init(&p->output_cond, NULL);
+
+ p->parent = fctx;
+ p->avctx = copy;
+
+ if (!copy) {
+ err = AVERROR(ENOMEM);
+ goto error;
+ }
+
+ *copy = *src;
+ copy->thread_opaque = p;
+ copy->pkt = &p->avpkt;
+
+ if (!i) {
+ src = copy;
+
+ if (codec->init)
+ err = codec->init(copy);
+
+ update_context_from_thread(avctx, copy, 1);
+ } else {
+ copy->priv_data = av_malloc(codec->priv_data_size);
+ if (!copy->priv_data) {
+ err = AVERROR(ENOMEM);
+ goto error;
+ }
+ memcpy(copy->priv_data, src->priv_data, codec->priv_data_size);
+ copy->internal = av_malloc(sizeof(AVCodecInternal));
+ if (!copy->internal) {
+ err = AVERROR(ENOMEM);
+ goto error;
+ }
+ *copy->internal = *src->internal;
+ copy->internal->is_copy = 1;
+
+ if (codec->init_thread_copy)
+ err = codec->init_thread_copy(copy);
+ }
+
+ if (err) goto error;
+
+ if (!pthread_create(&p->thread, NULL, frame_worker_thread, p))
+ p->thread_init = 1;
}
return 0;
+
+error:
+ frame_thread_free(avctx, i+1);
+
+ return err;
}
-int avcodec_thread_init(AVCodecContext *s, int thread_count){
+void ff_thread_flush(AVCodecContext *avctx)
+{
+ FrameThreadContext *fctx = avctx->thread_opaque;
+
+ if (!avctx->thread_opaque) return;
+
+ park_frame_worker_threads(fctx, avctx->thread_count);
+ if (fctx->prev_thread) {
+ if (fctx->prev_thread != &fctx->threads[0])
+ update_context_from_thread(fctx->threads[0].avctx, fctx->prev_thread->avctx, 0);
+ if (avctx->codec->flush)
+ avctx->codec->flush(fctx->threads[0].avctx);
+ }
+
+ fctx->next_decoding = fctx->next_finished = 0;
+ fctx->delaying = 1;
+ fctx->prev_thread = NULL;
+ for (int i = 0; i < avctx->thread_count; i++) {
+ PerThreadContext *p = &fctx->threads[i];
+ // Make sure decode flush calls with size=0 won't return old frames
+ p->got_frame = 0;
+
+ release_delayed_buffers(p);
+ }
+}
+
+static int *allocate_progress(PerThreadContext *p)
+{
int i;
- ThreadContext *c;
- WorkerContext *worker;
-
- s->thread_count= thread_count;
-
- assert(!s->thread_opaque);
- c= av_mallocz(sizeof(ThreadContext));
- worker= av_mallocz(sizeof(WorkerContext)*thread_count);
- s->thread_opaque= c;
- c->worker= worker;
-
- for(i=0; i<thread_count; i++){
-//printf("init semaphors %d\n", i); fflush(stdout);
- worker[i].avctx= s;
- if(sem_init(&worker[i].work_sem, 0, 0))
- goto fail;
- if(sem_init(&worker[i].done_sem, 0, 0))
- goto fail;
-//printf("create thread %d\n", i); fflush(stdout);
- if(pthread_create(&worker[i].thread, NULL, thread_func, &worker[i]))
- goto fail;
- }
-//printf("init done\n"); fflush(stdout);
-
- s->execute= avcodec_thread_execute;
+
+ for (i = 0; i < MAX_BUFFERS; i++)
+ if (!p->progress_used[i]) break;
+
+ if (i == MAX_BUFFERS) {
+ av_log(p->avctx, AV_LOG_ERROR, "allocate_progress() overflow\n");
+ return NULL;
+ }
+
+ p->progress_used[i] = 1;
+
+ return p->progress[i];
+}
+
+int ff_thread_get_buffer(AVCodecContext *avctx, AVFrame *f)
+{
+ PerThreadContext *p = avctx->thread_opaque;
+ int *progress, err;
+
+ f->owner = avctx;
+
+ if (!(avctx->active_thread_type&FF_THREAD_FRAME)) {
+ f->thread_opaque = NULL;
+ return avctx->get_buffer(avctx, f);
+ }
+
+ if (p->state != STATE_SETTING_UP &&
+ (avctx->codec->update_thread_context || !avctx->thread_safe_callbacks)) {
+ av_log(avctx, AV_LOG_ERROR, "get_buffer() cannot be called after ff_thread_finish_setup()\n");
+ return -1;
+ }
+
+ pthread_mutex_lock(&p->parent->buffer_mutex);
+ f->thread_opaque = progress = allocate_progress(p);
+
+ if (!progress) {
+ pthread_mutex_unlock(&p->parent->buffer_mutex);
+ return -1;
+ }
+
+ progress[0] =
+ progress[1] = -1;
+
+ if (avctx->thread_safe_callbacks ||
+ avctx->get_buffer == avcodec_default_get_buffer) {
+ err = avctx->get_buffer(avctx, f);
+ } else {
+ p->requested_frame = f;
+ p->state = STATE_GET_BUFFER;
+ pthread_mutex_lock(&p->progress_mutex);
+ pthread_cond_signal(&p->progress_cond);
+
+ while (p->state != STATE_SETTING_UP)
+ pthread_cond_wait(&p->progress_cond, &p->progress_mutex);
+
+ err = p->result;
+
+ pthread_mutex_unlock(&p->progress_mutex);
+
+ if (!avctx->codec->update_thread_context)
+ ff_thread_finish_setup(avctx);
+ }
+
+ if (err) {
+ free_progress(f);
+ f->thread_opaque = NULL;
+ }
+ pthread_mutex_unlock(&p->parent->buffer_mutex);
+
+ return err;
+}
+
+void ff_thread_release_buffer(AVCodecContext *avctx, AVFrame *f)
+{
+ PerThreadContext *p = avctx->thread_opaque;
+ FrameThreadContext *fctx;
+
+ if (!(avctx->active_thread_type&FF_THREAD_FRAME)) {
+ avctx->release_buffer(avctx, f);
+ return;
+ }
+
+ if (p->num_released_buffers >= MAX_BUFFERS) {
+ av_log(p->avctx, AV_LOG_ERROR, "too many thread_release_buffer calls!\n");
+ return;
+ }
+
+ if(avctx->debug & FF_DEBUG_BUFFERS)
+ av_log(avctx, AV_LOG_DEBUG, "thread_release_buffer called on pic %p\n", f);
+
+ fctx = p->parent;
+ pthread_mutex_lock(&fctx->buffer_mutex);
+ p->released_buffers[p->num_released_buffers++] = *f;
+ pthread_mutex_unlock(&fctx->buffer_mutex);
+ memset(f->data, 0, sizeof(f->data));
+}
+
+/**
+ * Set the threading algorithms used.
+ *
+ * Threading requires more than one thread.
+ * Frame threading requires entire frames to be passed to the codec,
+ * and introduces extra decoding delay, so is incompatible with low_delay.
+ *
+ * @param avctx The context.
+ */
+static void validate_thread_parameters(AVCodecContext *avctx)
+{
+ int frame_threading_supported = (avctx->codec->capabilities & CODEC_CAP_FRAME_THREADS)
+ && !(avctx->flags & CODEC_FLAG_TRUNCATED)
+ && !(avctx->flags & CODEC_FLAG_LOW_DELAY)
+ && !(avctx->flags2 & CODEC_FLAG2_CHUNKS);
+ if (avctx->thread_count == 1) {
+ avctx->active_thread_type = 0;
+ } else if (frame_threading_supported && (avctx->thread_type & FF_THREAD_FRAME)) {
+ avctx->active_thread_type = FF_THREAD_FRAME;
+ } else if (avctx->codec->capabilities & CODEC_CAP_SLICE_THREADS &&
+ avctx->thread_type & FF_THREAD_SLICE) {
+ avctx->active_thread_type = FF_THREAD_SLICE;
+ } else if (!(avctx->codec->capabilities & CODEC_CAP_AUTO_THREADS)) {
+ avctx->thread_count = 1;
+ avctx->active_thread_type = 0;
+ }
+}
+
+int ff_thread_init(AVCodecContext *avctx)
+{
+ if (avctx->thread_opaque) {
+ av_log(avctx, AV_LOG_ERROR, "avcodec_thread_init is ignored after avcodec_open\n");
+ return -1;
+ }
+
+#if HAVE_W32THREADS
+ w32thread_init();
+#endif
+
+ if (avctx->codec) {
+ validate_thread_parameters(avctx);
+
+ if (avctx->active_thread_type&FF_THREAD_SLICE)
+ return thread_init(avctx);
+ else if (avctx->active_thread_type&FF_THREAD_FRAME)
+ return frame_thread_init(avctx);
+ }
return 0;
-fail:
- avcodec_thread_free(s);
- return -1;
+}
+
+void ff_thread_free(AVCodecContext *avctx)
+{
+ if (avctx->active_thread_type&FF_THREAD_FRAME)
+ frame_thread_free(avctx, avctx->thread_count);
+ else
+ thread_free(avctx);
}
projects / ffmpeg / blobdiff