pipewire/src/examples/video-play-sync.c
Wim Taymans 9f2d873760 examples: set exclusive and reliable flags
We need exclusive port use if we negotiated SyncTimeline because there
can only be one consumer of the syncobj.

We also need to enable reliable transport if synctimeline is supported
but the release flag isn't.

Add some more logging to the port when the exclusive and reliable states
changed.

Fixes #4885
2025-10-14 11:56:22 +02:00

590 lines
17 KiB
C

/* PipeWire */
/* SPDX-FileCopyrightText: Copyright © 2025 Wim Taymans */
/* SPDX-License-Identifier: MIT */
/*
[title]
Video input stream using \ref pw_stream "pw_stream" and sync timeline.
[title]
*/
#include <stdio.h>
#include <unistd.h>
#include <signal.h>
#include <getopt.h>
#include <spa/utils/result.h>
#include <spa/param/video/format-utils.h>
#include <spa/param/tag-utils.h>
#include <spa/param/props.h>
#include <spa/param/latency-utils.h>
#include <spa/debug/format.h>
#include <spa/debug/pod.h>
#include <pipewire/pipewire.h>
#define WIDTH 1920
#define HEIGHT 1080
#define RATE 30
#define MAX_BUFFERS 64
#include "sdl.h"
struct pixel {
float r, g, b, a;
};
struct data {
const char *path;
SDL_Renderer *renderer;
SDL_Window *window;
SDL_Texture *texture;
struct pw_main_loop *loop;
struct pw_stream *stream;
struct spa_hook stream_listener;
struct spa_io_position *position;
struct spa_video_info format;
int32_t stride;
struct spa_rectangle size;
int counter;
SDL_Rect rect;
bool is_yuv;
bool with_synctimeline;
bool with_synctimeline_release;
bool force_synctimeline_release;
};
static void handle_events(struct data *data)
{
SDL_Event event;
while (SDL_PollEvent(&event)) {
switch (event.type) {
case SDL_QUIT:
pw_main_loop_quit(data->loop);
break;
}
}
}
/* our data processing function is in general:
*
* struct pw_buffer *b;
* b = pw_stream_dequeue_buffer(stream);
*
* .. do stuff with buffer ...
*
* pw_stream_queue_buffer(stream, b);
*/
static void
on_process(void *_data)
{
struct data *data = _data;
struct pw_stream *stream = data->stream;
struct pw_buffer *b;
struct spa_buffer *buf;
void *sdata, *ddata;
int sstride, dstride, ostride;
struct spa_meta_header *h;
struct spa_meta_sync_timeline *stl = NULL;
uint32_t i, j;
uint8_t *src, *dst;
uint64_t cmd;
b = NULL;
while (true) {
struct pw_buffer *t;
if ((t = pw_stream_dequeue_buffer(stream)) == NULL)
break;
if (b)
pw_stream_queue_buffer(stream, b);
b = t;
}
if (b == NULL) {
pw_log_warn("out of buffers: %m");
return;
}
buf = b->buffer;
pw_log_trace("new buffer %p", buf);
handle_events(data);
if ((sdata = buf->datas[0].data) == NULL)
goto done;
if ((h = spa_buffer_find_meta_data(buf, SPA_META_Header, sizeof(*h)))) {
uint64_t now = pw_stream_get_nsec(stream);
pw_log_debug("now:%"PRIu64" pts:%"PRIu64" diff:%"PRIi64,
now, h->pts, now - h->pts);
}
if ((stl = spa_buffer_find_meta_data(buf, SPA_META_SyncTimeline, sizeof(*stl))) &&
stl->acquire_point) {
/* wait before we can use the buffer */
if (read(buf->datas[1].fd, &cmd, sizeof(cmd)) < 0)
pw_log_warn("acquire_point wait error %m");
pw_log_debug("acquire_point:%"PRIu64, stl->acquire_point);
}
/* copy video image in texture */
if (data->is_yuv) {
void *datas[4];
sstride = data->stride;
if (buf->n_datas == 1) {
SDL_UpdateTexture(data->texture, NULL,
sdata, sstride);
} else {
datas[0] = sdata;
datas[1] = buf->datas[1].data;
datas[2] = buf->datas[2].data;
SDL_UpdateYUVTexture(data->texture, NULL,
datas[0], sstride,
datas[1], sstride / 2,
datas[2], sstride / 2);
}
}
else {
if (SDL_LockTexture(data->texture, NULL, &ddata, &dstride) < 0) {
fprintf(stderr, "Couldn't lock texture: %s\n", SDL_GetError());
}
sstride = buf->datas[0].chunk->stride;
if (sstride == 0)
sstride = buf->datas[0].chunk->size / data->size.height;
ostride = SPA_MIN(sstride, dstride);
src = sdata;
dst = ddata;
if (data->format.media_subtype == SPA_MEDIA_SUBTYPE_dsp) {
for (i = 0; i < data->size.height; i++) {
struct pixel *p = (struct pixel *) src;
for (j = 0; j < data->size.width; j++) {
dst[j * 4 + 0] = SPA_CLAMP((uint8_t)(p[j].r * 255.0f), 0u, 255u);
dst[j * 4 + 1] = SPA_CLAMP((uint8_t)(p[j].g * 255.0f), 0u, 255u);
dst[j * 4 + 2] = SPA_CLAMP((uint8_t)(p[j].b * 255.0f), 0u, 255u);
dst[j * 4 + 3] = SPA_CLAMP((uint8_t)(p[j].a * 255.0f), 0u, 255u);
}
src += sstride;
dst += dstride;
}
} else {
for (i = 0; i < data->size.height; i++) {
memcpy(dst, src, ostride);
src += sstride;
dst += dstride;
}
}
SDL_UnlockTexture(data->texture);
}
SDL_RenderClear(data->renderer);
/* now render the video */
SDL_RenderCopy(data->renderer, data->texture, &data->rect, NULL);
SDL_RenderPresent(data->renderer);
done:
pw_stream_queue_buffer(stream, b);
if (stl != NULL && stl->release_point) {
/* we promise to signal the release point */
if (data->with_synctimeline_release)
SPA_FLAG_CLEAR(stl->flags, SPA_META_SYNC_TIMELINE_UNSCHEDULED_RELEASE);
cmd = 1;
/* signal buffer release point */
write(buf->datas[2].fd, &cmd, sizeof(cmd));
pw_log_debug("release:%"PRIu64, stl->release_point);
}
}
static void on_stream_state_changed(void *_data, enum pw_stream_state old,
enum pw_stream_state state, const char *error)
{
struct data *data = _data;
fprintf(stderr, "stream state: \"%s\"\n", pw_stream_state_as_string(state));
switch (state) {
case PW_STREAM_STATE_UNCONNECTED:
pw_main_loop_quit(data->loop);
break;
case PW_STREAM_STATE_PAUSED:
/* because we started inactive, activate ourselves now */
pw_stream_set_active(data->stream, true);
break;
default:
break;
}
}
static void
on_stream_io_changed(void *_data, uint32_t id, void *area, uint32_t size)
{
struct data *data = _data;
switch (id) {
case SPA_IO_Position:
data->position = area;
break;
}
}
/* Be notified when the stream param changes. We're only looking at the
* format changes.
*
* We are now supposed to call pw_stream_finish_format() with success or
* failure, depending on if we can support the format. Because we gave
* a list of supported formats, this should be ok.
*
* As part of pw_stream_finish_format() we can provide parameters that
* will control the buffer memory allocation. This includes the metadata
* that we would like on our buffer, the size, alignment, etc.
*/
static void
on_stream_param_changed(void *_data, uint32_t id, const struct spa_pod *param)
{
struct data *data = _data;
struct pw_stream *stream = data->stream;
uint8_t params_buffer[1024];
struct spa_pod_builder b = SPA_POD_BUILDER_INIT(params_buffer, sizeof(params_buffer));
struct spa_pod_frame f;
const struct spa_pod *params[5];
uint32_t n_params = 0;
Uint32 sdl_format;
void *d;
int32_t mult, size, blocks;
if (param != NULL && id == SPA_PARAM_Tag) {
spa_debug_pod(0, NULL, param);
return;
}
if (param != NULL && id == SPA_PARAM_Latency) {
struct spa_latency_info info;
if (spa_latency_parse(param, &info) >= 0)
fprintf(stderr, "got latency: %"PRIu64"\n", (info.min_ns + info.max_ns) / 2);
return;
}
/* NULL means to clear the format */
if (param == NULL || id != SPA_PARAM_Format)
return;
fprintf(stderr, "got format:\n");
spa_debug_format(2, NULL, param);
if (spa_format_parse(param, &data->format.media_type, &data->format.media_subtype) < 0)
return;
if (data->format.media_type != SPA_MEDIA_TYPE_video)
return;
switch (data->format.media_subtype) {
case SPA_MEDIA_SUBTYPE_raw:
/* call a helper function to parse the format for us. */
spa_format_video_raw_parse(param, &data->format.info.raw);
sdl_format = id_to_sdl_format(data->format.info.raw.format);
data->size = SPA_RECTANGLE(data->format.info.raw.size.width,
data->format.info.raw.size.height);
mult = 1;
break;
case SPA_MEDIA_SUBTYPE_dsp:
spa_format_video_dsp_parse(param, &data->format.info.dsp);
if (data->format.info.dsp.format != SPA_VIDEO_FORMAT_DSP_F32)
return;
sdl_format = SDL_PIXELFORMAT_RGBA32;
data->size = SPA_RECTANGLE(data->position->video.size.width,
data->position->video.size.height);
mult = 4;
break;
default:
sdl_format = SDL_PIXELFORMAT_UNKNOWN;
break;
}
if (sdl_format == SDL_PIXELFORMAT_UNKNOWN) {
pw_stream_set_error(stream, -EINVAL, "unknown pixel format");
return;
}
if (data->size.width == 0 || data->size.height == 0) {
pw_stream_set_error(stream, -EINVAL, "invalid size");
return;
}
data->texture = SDL_CreateTexture(data->renderer,
sdl_format,
SDL_TEXTUREACCESS_STREAMING,
data->size.width,
data->size.height);
switch(sdl_format) {
case SDL_PIXELFORMAT_YV12:
case SDL_PIXELFORMAT_IYUV:
data->stride = data->size.width;
size = (data->stride * data->size.height) * 3 / 2;
data->is_yuv = true;
blocks = 3;
break;
case SDL_PIXELFORMAT_YUY2:
data->is_yuv = true;
data->stride = data->size.width * 2;
size = (data->stride * data->size.height);
blocks = 1;
break;
default:
if (SDL_LockTexture(data->texture, NULL, &d, &data->stride) < 0) {
fprintf(stderr, "Couldn't lock texture: %s\n", SDL_GetError());
data->stride = data->size.width * 2;
} else
SDL_UnlockTexture(data->texture);
size = data->stride * data->size.height;
blocks = 1;
break;
}
data->rect.x = 0;
data->rect.y = 0;
data->rect.w = data->size.width;
data->rect.h = data->size.height;
if (data->with_synctimeline) {
/* first add Buffer with 3 blocks (1 data, 2 sync fds). */
spa_pod_builder_push_object(&b, &f, SPA_TYPE_OBJECT_ParamBuffers, SPA_PARAM_Buffers);
spa_pod_builder_add(&b,
SPA_PARAM_BUFFERS_buffers, SPA_POD_CHOICE_RANGE_Int(8, 2, MAX_BUFFERS),
SPA_PARAM_BUFFERS_blocks, SPA_POD_Int(3),
SPA_PARAM_BUFFERS_size, SPA_POD_Int(size * mult),
SPA_PARAM_BUFFERS_stride, SPA_POD_Int(data->stride * mult),
SPA_PARAM_BUFFERS_dataType, SPA_POD_CHOICE_FLAGS_Int((1<<SPA_DATA_MemFd)),
0);
/* this depends on the negotiation of the SyncTimeline metadata */
spa_pod_builder_prop(&b, SPA_PARAM_BUFFERS_metaType, SPA_POD_PROP_FLAG_MANDATORY);
spa_pod_builder_int(&b, 1<<SPA_META_SyncTimeline);
params[n_params++] = spa_pod_builder_pop(&b, &f);
/* explicit sync information */
spa_pod_builder_push_object(&b, &f, SPA_TYPE_OBJECT_ParamMeta, SPA_PARAM_Meta);
spa_pod_builder_add(&b,
SPA_PARAM_META_type, SPA_POD_Id(SPA_META_SyncTimeline),
SPA_PARAM_META_size, SPA_POD_Int(sizeof(struct spa_meta_sync_timeline)),
0);
if (data->with_synctimeline_release) {
uint32_t flags = data->force_synctimeline_release ?
/* both sides need compatible features */
SPA_POD_PROP_FLAG_MANDATORY :
/* drop features flags if not provided by both sides */
SPA_POD_PROP_FLAG_DROP;
spa_pod_builder_prop(&b, SPA_PARAM_META_features, flags);
spa_pod_builder_int(&b, SPA_META_FEATURE_SYNC_TIMELINE_RELEASE);
}
params[n_params++] = spa_pod_builder_pop(&b, &f);
}
/* fallback for when the synctimeline is not negotiated */
params[n_params++] = spa_pod_builder_add_object(&b,
SPA_TYPE_OBJECT_ParamBuffers, SPA_PARAM_Buffers,
SPA_PARAM_BUFFERS_buffers, SPA_POD_CHOICE_RANGE_Int(8, 2, MAX_BUFFERS),
SPA_PARAM_BUFFERS_blocks, SPA_POD_Int(blocks),
SPA_PARAM_BUFFERS_size, SPA_POD_Int(size * mult),
SPA_PARAM_BUFFERS_stride, SPA_POD_Int(data->stride * mult),
SPA_PARAM_BUFFERS_dataType, SPA_POD_CHOICE_FLAGS_Int((1<<SPA_DATA_MemFd)));
/* a header metadata with timing information */
params[n_params++] = spa_pod_builder_add_object(&b,
SPA_TYPE_OBJECT_ParamMeta, SPA_PARAM_Meta,
SPA_PARAM_META_type, SPA_POD_Id(SPA_META_Header),
SPA_PARAM_META_size, SPA_POD_Int(sizeof(struct spa_meta_header)));
/* we are done */
pw_stream_update_params(stream, params, n_params);
}
/* these are the stream events we listen for */
static const struct pw_stream_events stream_events = {
PW_VERSION_STREAM_EVENTS,
.state_changed = on_stream_state_changed,
.io_changed = on_stream_io_changed,
.param_changed = on_stream_param_changed,
.process = on_process,
};
static int build_format(struct data *data, struct spa_pod_builder *b, const struct spa_pod **params)
{
uint32_t n_params = 0;
SDL_RendererInfo info;
SDL_GetRendererInfo(data->renderer, &info);
params[n_params++] = sdl_build_formats(&info, b);
fprintf(stderr, "supported SDL formats:\n");
spa_debug_format(2, NULL, params[0]);
params[n_params++] = spa_pod_builder_add_object(b,
SPA_TYPE_OBJECT_Format, SPA_PARAM_EnumFormat,
SPA_FORMAT_mediaType, SPA_POD_Id(SPA_MEDIA_TYPE_video),
SPA_FORMAT_mediaSubtype, SPA_POD_Id(SPA_MEDIA_SUBTYPE_dsp),
SPA_FORMAT_VIDEO_format, SPA_POD_Id(SPA_VIDEO_FORMAT_DSP_F32));
fprintf(stderr, "supported DSP formats:\n");
spa_debug_format(2, NULL, params[1]);
return n_params;
}
static void do_quit(void *userdata, int signal_number)
{
struct data *data = userdata;
pw_main_loop_quit(data->loop);
}
static void show_help(struct data *data, const char *name, bool is_error)
{
FILE *fp;
fp = is_error ? stderr : stdout;
fprintf(fp,
"%s [options]\n"
" -h, --help Show this help\n"
" --version Show version\n"
" -r, --remote Remote daemon name\n"
" -S, --sync Enable SyncTimeline\n"
" -R, --release Enable RELEASE feature\n"
" -F, --force-release RELEASE feature needs to be present\n"
"\n", name);
}
int main(int argc, char *argv[])
{
struct data data = { 0, };
const struct spa_pod *params[3];
uint8_t buffer[1024];
struct spa_pod_builder b = SPA_POD_BUILDER_INIT(buffer, sizeof(buffer));
struct pw_properties *props;
int res, n_params;
static const struct option long_options[] = {
{ "help", no_argument, NULL, 'h' },
{ "version", no_argument, NULL, 'V' },
{ "remote", required_argument, NULL, 'r' },
{ "sync", no_argument, NULL, 'S' },
{ "release", no_argument, NULL, 'R' },
{ "force-release", no_argument, NULL, 'F' },
{ NULL, 0, NULL, 0}
};
char *opt_remote = NULL;
int c;
pw_init(&argc, &argv);
while ((c = getopt_long(argc, argv, "hVr:SRF", long_options, NULL)) != -1) {
switch (c) {
case 'h':
show_help(&data, argv[0], false);
return 0;
case 'V':
printf("%s\n"
"Compiled with libpipewire %s\n"
"Linked with libpipewire %s\n",
argv[0],
pw_get_headers_version(),
pw_get_library_version());
return 0;
case 'r':
opt_remote = optarg;
break;
case 'F':
data.force_synctimeline_release = true;
SPA_FALLTHROUGH;
case 'R':
data.with_synctimeline_release = true;
SPA_FALLTHROUGH;
case 'S':
data.with_synctimeline = true;
break;
default:
show_help(&data, argv[0], true);
return -1;
}
}
/* create a main loop */
data.loop = pw_main_loop_new(NULL);
pw_loop_add_signal(pw_main_loop_get_loop(data.loop), SIGINT, do_quit, &data);
pw_loop_add_signal(pw_main_loop_get_loop(data.loop), SIGTERM, do_quit, &data);
/* create a simple stream, the simple stream manages to core and remote
* objects for you if you don't need to deal with them
*
* If you plan to autoconnect your stream, you need to provide at least
* media, category and role properties
*
* Pass your events and a user_data pointer as the last arguments. This
* will inform you about the stream state. The most important event
* you need to listen to is the process event where you need to consume
* the data provided to you.
*/
props = pw_properties_new(PW_KEY_MEDIA_TYPE, "Video",
PW_KEY_MEDIA_CATEGORY, "Capture",
PW_KEY_MEDIA_ROLE, "Camera",
PW_KEY_REMOTE_NAME, opt_remote,
NULL),
data.path = optind < argc ? argv[optind++] : "video-src-sync";
if (data.path)
pw_properties_set(props, PW_KEY_TARGET_OBJECT, data.path);
data.stream = pw_stream_new_simple(
pw_main_loop_get_loop(data.loop),
"video-play",
props,
&stream_events,
&data);
if (SDL_Init(SDL_INIT_VIDEO) < 0) {
fprintf(stderr, "can't initialize SDL: %s\n", SDL_GetError());
return -1;
}
if (SDL_CreateWindowAndRenderer
(WIDTH, HEIGHT, SDL_WINDOW_RESIZABLE, &data.window, &data.renderer)) {
fprintf(stderr, "can't create window: %s\n", SDL_GetError());
return -1;
}
/* build the extra parameters to connect with. To connect, we can provide
* a list of supported formats. We use a builder that writes the param
* object to the stack. */
n_params = build_format(&data, &b, params);
/* now connect the stream, we need a direction (input/output),
* an optional target node to connect to, some flags and parameters
*/
if ((res = pw_stream_connect(data.stream,
PW_DIRECTION_INPUT,
PW_ID_ANY,
PW_STREAM_FLAG_AUTOCONNECT | /* try to automatically connect this stream */
PW_STREAM_FLAG_INACTIVE | /* we will activate ourselves */
PW_STREAM_FLAG_MAP_BUFFERS, /* mmap the buffer data for us */
params, n_params)) /* extra parameters, see above */ < 0) {
fprintf(stderr, "can't connect: %s\n", spa_strerror(res));
return -1;
}
/* do things until we quit the mainloop */
pw_main_loop_run(data.loop);
pw_stream_destroy(data.stream);
pw_main_loop_destroy(data.loop);
SDL_DestroyTexture(data.texture);
SDL_DestroyRenderer(data.renderer);
SDL_DestroyWindow(data.window);
pw_deinit();
return 0;
}