graph: use spa_node as implementation

Always use a spa_node as the graph implementation, implementing the
methods is just as easy.
Plug some mem leaks in remote

spa/include/spa/graph-scheduler1.h

@@ -45,24 +45,6 @@ static inline void spa_graph_data_init(struct spa_graph_data *data,
 	data->node = NULL;
 }
 
-static inline int spa_graph_node_impl_input(void *data)
-{
-	struct spa_node *n = data;
-	return spa_node_process_input(n);
-}
-
-static inline int spa_graph_node_impl_output(void *data)
-{
-	struct spa_node *n = data;
-	return spa_node_process_output(n);
-}
-
-static const struct spa_graph_node_callbacks spa_graph_node_impl_default = {
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
-	spa_graph_node_impl_input,
-	spa_graph_node_impl_output,
-};
-
 static inline void spa_graph_data_port_check(struct spa_graph_data *data, struct spa_graph_port *port)
 {
 	struct spa_graph_node *node = port->node;
@@ -70,7 +52,8 @@ static inline void spa_graph_data_port_check(struct spa_graph_data *data, struct
 	if (port->io->status == SPA_RESULT_HAVE_BUFFER)
 		node->ready_in++;
 
-	debug("port %p node %p check %d %d %d\n", port, node, port->io->status, node->ready_in, node->required_in);
+	debug("port %p node %p check %d %d %d\n", port, node,
+	      port->io->status, node->ready_in, node->required_in);
 
 	if (node->required_in > 0 && node->ready_in == node->required_in) {
 		node->state = SPA_GRAPH_STATE_IN;
@@ -100,7 +83,7 @@ static inline bool spa_graph_data_iterate(struct spa_graph_data *data)
 
 		switch (n->state) {
 		case SPA_GRAPH_STATE_IN:
-			state = n->callbacks->process_input(n->callbacks_data);
+			state = spa_node_process_input(n->implementation);
 			if (state == SPA_RESULT_NEED_BUFFER)
 				n->state = SPA_GRAPH_STATE_CHECK_IN;
 			else if (state == SPA_RESULT_HAVE_BUFFER)
@@ -112,7 +95,7 @@ static inline bool spa_graph_data_iterate(struct spa_graph_data *data)
 			break;
 
 		case SPA_GRAPH_STATE_OUT:
-			state = n->callbacks->process_output(n->callbacks_data);
+			state = spa_node_process_output(n->implementation);
 			if (state == SPA_RESULT_NEED_BUFFER)
 				n->state = SPA_GRAPH_STATE_CHECK_IN;
 			else if (state == SPA_RESULT_HAVE_BUFFER)

spa/include/spa/graph-scheduler3.h

@@ -26,38 +26,6 @@ extern "C" {
 
 #include <spa/graph.h>
 
-static inline int spa_graph_node_impl_input(void *data)
-{
-	struct spa_node *n = data;
-	return spa_node_process_input(n);
-}
-
-static inline int spa_graph_node_impl_output(void *data)
-{
-	struct spa_node *n = data;
-	return spa_node_process_output(n);
-}
-
-static const struct spa_graph_node_callbacks spa_graph_node_impl_default = {
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
-	spa_graph_node_impl_input,
-	spa_graph_node_impl_output,
-};
-
-static inline int spa_graph_port_impl_reuse_buffer(void *data,
-						   uint32_t buffer_id)
-{
-	struct spa_graph_port *port = data;
-	struct spa_node *node = port->node->callbacks_data;
-	debug("port %p reuse buffer %d\n", port, buffer_id);
-	return spa_node_port_reuse_buffer(node, port->port_id, buffer_id);
-}
-
-static const struct spa_graph_port_callbacks spa_graph_port_impl_default = {
-	SPA_VERSION_GRAPH_PORT_CALLBACKS,
-	spa_graph_port_impl_reuse_buffer,
-};
-
 static inline int spa_graph_impl_need_input(void *data, struct spa_graph_node *node)
 {
 	struct spa_graph_port *p;
@@ -85,7 +53,7 @@ static inline int spa_graph_impl_need_input(void *data, struct spa_graph_node *n
 	}
 
 	spa_list_for_each_safe(n, t, &ready, ready_link) {
-		n->state = n->callbacks->process_output(n->callbacks_data);
+		n->state = spa_node_process_output(n->implementation);
 		debug("peer %p processed out %d\n", n, n->state);
 		if (n->state == SPA_RESULT_NEED_BUFFER)
 			spa_graph_need_input(n->graph, n);
@@ -102,7 +70,7 @@ static inline int spa_graph_impl_need_input(void *data, struct spa_graph_node *n
 	debug("node %p ready_in:%d required_in:%d\n", node, node->ready_in, node->required_in);
 
 	if (node->required_in > 0 && node->ready_in == node->required_in) {
-		node->state = node->callbacks->process_input(node->callbacks_data);
+		node->state = spa_node_process_input(node->implementation);
 		debug("node %p processed in %d\n", node, node->state);
 		if (node->state == SPA_RESULT_HAVE_BUFFER) {
 			spa_list_for_each(p, &node->ports[SPA_DIRECTION_OUTPUT], link) {
@@ -143,7 +111,7 @@ static inline int spa_graph_impl_have_output(void *data, struct spa_graph_node *
 	}
 
 	spa_list_for_each_safe(n, t, &ready, ready_link) {
-		n->state = n->callbacks->process_input(n->callbacks_data);
+		n->state = spa_node_process_input(n->implementation);
 		debug("node %p chain processed in %d\n", n, n->state);
 		if (n->state == SPA_RESULT_HAVE_BUFFER)
 			spa_graph_have_output(n->graph, n);
@@ -159,7 +127,7 @@ static inline int spa_graph_impl_have_output(void *data, struct spa_graph_node *
 		n->ready_link.next = NULL;
 	}
 
-	node->state = node->callbacks->process_output(node->callbacks_data);
+	node->state = spa_node_process_output(node->implementation);
 	debug("node %p processed out %d\n", node, node->state);
 	if (node->state == SPA_RESULT_NEED_BUFFER) {
 		node->ready_in = 0;

spa/include/spa/graph.h

@@ -56,21 +56,7 @@ struct spa_graph {
 
 #define spa_graph_need_input(g,n)	((g)->callbacks->need_input((g)->callbacks_data, (n)))
 #define spa_graph_have_output(g,n)	((g)->callbacks->have_output((g)->callbacks_data, (n)))
-
+#define spa_graph_reuse_buffer(g,n,p,i)	((g)->callbacks->reuse_buffer((g)->callbacks_data, (n),(p),(i)))
-struct spa_graph_node_callbacks {
-#define SPA_VERSION_GRAPH_NODE_CALLBACKS	0
-	uint32_t version;
-
-	int (*process_input) (void *data);
-	int (*process_output) (void *data);
-};
-
-struct spa_graph_port_callbacks {
-#define SPA_VERSION_GRAPH_PORT_CALLBACKS	0
-	uint32_t version;
-
-	int (*reuse_buffer) (void *data, uint32_t buffer_id);
-};
 
 struct spa_graph_node {
 	struct spa_list link;		/**< link in graph nodes list */
@@ -82,9 +68,7 @@ struct spa_graph_node {
 	uint32_t required_in;		/**< required number of ports */
 	uint32_t ready_in;		/**< number of ports with data */
 	int state;			/**< state of the node */
-	/** callbacks and data */
+	struct spa_node *implementation;/**< node implementation */
-	const struct spa_graph_node_callbacks *callbacks;
-	void *callbacks_data;
 	void *scheduler_data;		/**< scheduler private data */
 };
 
@@ -96,9 +80,6 @@ struct spa_graph_port {
 	uint32_t flags;			/**< port flags */
 	struct spa_port_io *io;		/**< io area of the port */
 	struct spa_graph_port *peer;	/**< peer */
-	/** callbacks and data */
-	const struct spa_graph_port_callbacks *callbacks;
-	void *callbacks_data;
 	void *scheduler_data;		/**< scheduler private data */
 };
 
@@ -127,12 +108,10 @@ spa_graph_node_init(struct spa_graph_node *node)
 }
 
 static inline void
-spa_graph_node_set_callbacks(struct spa_graph_node *node,
+spa_graph_node_set_implementation(struct spa_graph_node *node,
-			     const struct spa_graph_node_callbacks *callbacks,
+				  struct spa_node *implementation)
-			     void *data)
 {
-	node->callbacks = callbacks;
+	node->implementation = implementation;
-	node->callbacks_data = data;
 }
 
 static inline void
@@ -160,15 +139,6 @@ spa_graph_port_init(struct spa_graph_port *port,
 	port->io = io;
 }
 
-static inline void
-spa_graph_port_set_callbacks(struct spa_graph_port *port,
-			     const struct spa_graph_port_callbacks *callbacks,
-			     void *data)
-{
-	port->callbacks = callbacks;
-	port->callbacks_data = data;
-}
-
 static inline void
 spa_graph_port_add(struct spa_graph_node *node,
 		   struct spa_graph_port *port)

spa/tests/test-graph.c

@@ -337,13 +337,13 @@ static int make_nodes(struct data *data, const char *device)
 	spa_node_port_set_io(data->sink, SPA_DIRECTION_INPUT, 0, &data->volume_sink_io[0]);
 
 	spa_graph_node_init(&data->source_node);
-	spa_graph_node_set_callbacks(&data->source_node, &spa_graph_node_impl_default, data->source);
+	spa_graph_node_set_implementation(&data->source_node, data->source);
 	spa_graph_node_add(&data->graph, &data->source_node);
 	spa_graph_port_init(&data->source_out, SPA_DIRECTION_OUTPUT, 0, 0, &data->source_volume_io[0]);
 	spa_graph_port_add(&data->source_node, &data->source_out);
 
 	spa_graph_node_init(&data->volume_node);
-	spa_graph_node_set_callbacks(&data->volume_node, &spa_graph_node_impl_default, data->volume);
+	spa_graph_node_set_implementation(&data->volume_node, data->volume);
 	spa_graph_node_add(&data->graph, &data->volume_node);
 	spa_graph_port_init(&data->volume_in, SPA_DIRECTION_INPUT, 0, 0, &data->source_volume_io[0]);
 	spa_graph_port_add(&data->volume_node, &data->volume_in);
@@ -354,7 +354,7 @@ static int make_nodes(struct data *data, const char *device)
 	spa_graph_port_add(&data->volume_node, &data->volume_out);
 
 	spa_graph_node_init(&data->sink_node);
-	spa_graph_node_set_callbacks(&data->sink_node, &spa_graph_node_impl_default, data->sink);
+	spa_graph_node_set_implementation(&data->sink_node, data->sink);
 	spa_graph_node_add(&data->graph, &data->sink_node);
 	spa_graph_port_init(&data->sink_in, SPA_DIRECTION_INPUT, 0, 0, &data->volume_sink_io[0]);
 	spa_graph_port_add(&data->sink_node, &data->sink_in);

spa/tests/test-mixer.c

@@ -414,19 +414,19 @@ static int make_nodes(struct data *data, const char *device)
 
 #ifdef USE_GRAPH
 	spa_graph_node_init(&data->source1_node);
-	spa_graph_node_set_callbacks(&data->source1_node, &spa_graph_node_impl_default, data->source1);
+	spa_graph_node_set_implementation(&data->source1_node, data->source1);
 	spa_graph_port_init(&data->source1_out, SPA_DIRECTION_OUTPUT, 0, 0, &data->source1_mix_io[0]);
 	spa_graph_port_add(&data->source1_node, &data->source1_out);
 	spa_graph_node_add(&data->graph, &data->source1_node);
 
 	spa_graph_node_init(&data->source2_node);
-	spa_graph_node_set_callbacks(&data->source2_node, &spa_graph_node_impl_default, data->source2);
+	spa_graph_node_set_implementation(&data->source2_node, data->source2);
 	spa_graph_port_init(&data->source2_out, SPA_DIRECTION_OUTPUT, 0, 0, &data->source2_mix_io[0]);
 	spa_graph_port_add(&data->source2_node, &data->source2_out);
 	spa_graph_node_add(&data->graph, &data->source2_node);
 
 	spa_graph_node_init(&data->mix_node);
-	spa_graph_node_set_callbacks(&data->mix_node, &spa_graph_node_impl_default, data->mix);
+	spa_graph_node_set_implementation(&data->mix_node, data->mix);
 	spa_graph_port_init(&data->mix_in[0], SPA_DIRECTION_INPUT,
 			    data->mix_ports[0], 0, &data->source1_mix_io[0]);
 	spa_graph_port_add(&data->mix_node, &data->mix_in[0]);
@@ -442,7 +442,7 @@ static int make_nodes(struct data *data, const char *device)
 	spa_graph_port_add(&data->mix_node, &data->mix_out);
 
 	spa_graph_node_init(&data->sink_node);
-	spa_graph_node_set_callbacks(&data->sink_node, &spa_graph_node_impl_default, data->sink);
+	spa_graph_node_set_implementation(&data->sink_node, data->sink);
 	spa_graph_port_init(&data->sink_in, SPA_DIRECTION_INPUT, 0, 0, &data->mix_sink_io[0]);
 	spa_graph_port_add(&data->sink_node, &data->sink_in);
 	spa_graph_node_add(&data->graph, &data->sink_node);

spa/tests/test-perf.c

@@ -362,7 +362,7 @@ static int make_nodes(struct data *data)
 	spa_node_port_set_io(data->sink, SPA_DIRECTION_INPUT, 0, &data->source_sink_io[0]);
 
 	spa_graph_node_init(&data->source_node);
-	spa_graph_node_set_callbacks(&data->source_node, &spa_graph_node_impl_default, data->source);
+	spa_graph_node_set_implementation(&data->source_node, data->source);
 	spa_graph_node_add(&data->graph, &data->source_node);
 
 	data->source_node.flags = (data->mode & MODE_ASYNC_PUSH) ? SPA_GRAPH_NODE_FLAG_ASYNC : 0;
@@ -370,7 +370,7 @@ static int make_nodes(struct data *data)
 	spa_graph_port_add(&data->source_node, &data->source_out);
 
 	spa_graph_node_init(&data->sink_node);
-	spa_graph_node_set_callbacks(&data->sink_node, &spa_graph_node_impl_default, data->sink);
+	spa_graph_node_set_implementation(&data->sink_node, data->sink);
 	spa_graph_node_add(&data->graph, &data->sink_node);
 
 	data->sink_node.flags = (data->mode & MODE_ASYNC_PULL) ? SPA_GRAPH_NODE_FLAG_ASYNC : 0;

src/examples/export-sink.c

@@ -448,7 +448,7 @@ static void make_node(struct data *data)
 	struct pw_properties *props;
 
 	props = pw_properties_new(
-			//"pipewire.target.node", port_path,
+			"pipewire.target.node", data->path,
 			"pipewire.autoconnect", "1",
 			NULL);
 

src/examples/export-v4l2.c

@@ -19,6 +19,7 @@
 
 #include <stdio.h>
 #include <sys/mman.h>
+#include <signal.h>
 
 #include <spa/type-map.h>
 #include <spa/format-utils.h>
@@ -79,8 +80,6 @@ static void make_node(struct data *data)
                                   "spa.factory.name", "v4l2-source", NULL);
         data->node = pw_node_factory_create_node(factory, NULL, "v4l2-source", props);
 
-	pw_node_register(data->node);
-
 	pw_remote_export(data->remote, data->node);
 }
 
@@ -110,6 +109,12 @@ static const struct pw_remote_events remote_events = {
 	.state_changed = on_state_changed,
 };
 
+static void do_quit(void *data, int signal_number)
+{
+        struct data *d = data;
+	d->running = false;
+}
+
 int main(int argc, char *argv[])
 {
 	struct data data = { 0, };
@@ -117,6 +122,9 @@ int main(int argc, char *argv[])
 	pw_init(&argc, &argv);
 
 	data.loop = pw_loop_new(NULL);
+        pw_loop_add_signal(data.loop, SIGINT, do_quit, &data);
+        pw_loop_add_signal(data.loop, SIGTERM, do_quit, &data);
+
 	data.running = true;
 	data.core = pw_core_new(data.loop, NULL);
 	data.t = pw_core_get_type(data.core);
@@ -134,10 +142,14 @@ int main(int argc, char *argv[])
 
 	pw_loop_enter(data.loop);
 	while (data.running) {
-		pw_loop_iterate(data.loop, -1);
+		pw_loop_iterate(data.loop, 100);
 	}
 	pw_loop_leave(data.loop);
 
+	pw_remote_destroy(data.remote);
+	if (data.node)
+		pw_node_destroy(data.node);
+	pw_core_destroy(data.core);
 	pw_loop_destroy(data.loop);
 
 	return 0;

src/examples/local-v4l2.c

@@ -107,7 +107,7 @@ static void handle_events(struct data *data)
 	while (SDL_PollEvent(&event)) {
 		switch (event.type) {
 		case SDL_QUIT:
-			exit(0);
+			data->running = false;
 			break;
 		}
 	}
@@ -487,10 +487,13 @@ int main(int argc, char *argv[])
 
 	pw_loop_enter(data.loop);
 	while (data.running) {
-		pw_loop_iterate(data.loop, -1);
+		pw_loop_iterate(data.loop, 100);
 	}
 	pw_loop_leave(data.loop);
 
+	pw_link_destroy(data.link);
+	pw_node_destroy(data.node);
+	pw_core_destroy(data.core);
 	pw_loop_destroy(data.loop);
 
 	return 0;

src/modules/module-jack.c

@@ -1124,7 +1124,7 @@ static void jack_node_pull(void *data)
 	spa_list_for_each(p, &n->ports[SPA_DIRECTION_INPUT], link) {
 		if ((pp = p->peer) == NULL || ((pn = pp->node) == NULL))
 			continue;
-		pn->state = pn->callbacks->process_input(pn->callbacks_data);
+		pn->state = spa_node_process_input(pn->implementation);
 	}
 }
 
@@ -1154,7 +1154,7 @@ static void jack_node_push(void *data)
 	spa_list_for_each(p, &n->ports[SPA_DIRECTION_INPUT], link) {
 		if ((pp = p->peer) == NULL || ((pn = pp->node) == NULL))
 			continue;
-		pn->state = pn->callbacks->process_output(pn->callbacks_data);
+		pn->state = spa_node_process_output(pn->implementation);
 	}
 
 	spa_list_for_each(node, &impl->rt.nodes, graph_link) {
@@ -1163,25 +1163,25 @@ static void jack_node_push(void *data)
 		spa_list_for_each(p, &n->ports[SPA_DIRECTION_OUTPUT], link) {
 			if ((pp = p->peer) == NULL || ((pn = pp->node) == NULL))
 				continue;
-			pn->state = pn->callbacks->process_output(pn->callbacks_data);
+			pn->state = spa_node_process_output(pn->implementation);
 		}
-		n->state = n->callbacks->process_output(n->callbacks_data);
+		n->state = spa_node_process_output(n->implementation);
 
 		/* mix inputs */
 		spa_list_for_each(p, &n->ports[SPA_DIRECTION_INPUT], link) {
 			if ((pp = p->peer) == NULL || ((pn = pp->node) == NULL))
 				continue;
-			pn->state = pn->callbacks->process_output(pn->callbacks_data);
+			pn->state = spa_node_process_output(pn->implementation);
-			pn->state = pn->callbacks->process_input(pn->callbacks_data);
+			pn->state = spa_node_process_input(pn->implementation);
 		}
 
-		n->state = n->callbacks->process_input(n->callbacks_data);
+		n->state = spa_node_process_input(n->implementation);
 
 		/* tee outputs */
 		spa_list_for_each(p, &n->ports[SPA_DIRECTION_OUTPUT], link) {
 			if ((pp = p->peer) == NULL || ((pn = pp->node) == NULL))
 				continue;
-			pn->state = pn->callbacks->process_input(pn->callbacks_data);
+			pn->state = spa_node_process_input(pn->implementation);
 		}
 	}
 

src/modules/module-jack/jack-node.c

@@ -100,6 +100,8 @@ struct port_data {
 
 	bool driver_port;
 
+	struct spa_node mix_node;
+
 	struct spa_port_info info;
 
 	struct spa_port_io *io;
@@ -278,7 +280,7 @@ static int driver_process_output(struct spa_node *node)
 	spa_hook_list_call(&nd->listener_list, struct pw_jack_node_events, pull);
 
 	spa_list_for_each(p, &gn->ports[SPA_DIRECTION_INPUT], link) {
-		struct pw_port *port = p->callbacks_data;
+		struct pw_port *port = p->scheduler_data;
 		struct port_data *ipd = pw_port_get_user_data(port);
 		struct spa_port_io *in_io = ipd->io;
 		struct buffer *in;
@@ -327,7 +329,7 @@ static int node_process_input(struct spa_node *node)
 				     &server->synchro_table[ref_num]);
 
 	spa_list_for_each(p, &gn->ports[SPA_DIRECTION_OUTPUT], link) {
-		struct pw_port *port = p->callbacks_data;
+		struct pw_port *port = p->scheduler_data;
 		struct port_data *opd = pw_port_get_user_data(port);
 		struct spa_port_io *out_io = opd->io;
 		out_io->buffer_id = 0;
@@ -346,7 +348,7 @@ static int node_process_output(struct spa_node *node)
 
 	pw_log_trace(NAME " %p: process output", nd);
 	spa_list_for_each(p, &gn->ports[SPA_DIRECTION_INPUT], link) {
-		struct pw_port *port = p->callbacks_data;
+		struct pw_port *port = p->scheduler_data;
 		struct port_data *ipd = pw_port_get_user_data(port);
 		struct spa_port_io *in_io = ipd->io;
 		in_io->buffer_id = 0;
@@ -580,18 +582,18 @@ static const struct spa_node node_impl = {
 	.process_output = node_process_output,
 };
 
-static int schedule_mix_input(void *data)
+static int schedule_mix_input(struct spa_node *_node)
 {
-	struct pw_jack_port *this = data;
+	struct port_data *pd = SPA_CONTAINER_OF(_node, struct port_data, mix_node);
+	struct pw_jack_port *this = &pd->port;
 	struct spa_graph_node *node = &this->port->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->port->rt.mix_port.io;
-	struct port_data *pd = SPA_CONTAINER_OF(this, struct port_data, port);
 	size_t buffer_size = pd->node->node.server->engine_control->buffer_size;
 	int layer = 0;
 
 	spa_list_for_each(p, &node->ports[SPA_DIRECTION_INPUT], link) {
-		struct pw_link *link = p->callbacks_data;
+		struct pw_link *link = p->scheduler_data;
 		struct spa_buffer *inbuf;
 
 		pw_log_trace("mix %p: input %d %d", node, p->io->buffer_id, link->output->n_buffers);
@@ -616,9 +618,10 @@ static int schedule_mix_input(void *data)
 	return SPA_RESULT_HAVE_BUFFER;
 }
 
-static int schedule_mix_output(void *data)
+static int schedule_mix_output(struct spa_node *_node)
 {
-	struct pw_jack_port *this = data;
+	struct port_data *pd = SPA_CONTAINER_OF(_node, struct port_data, mix_node);
+	struct pw_jack_port *this = &pd->port;
 	struct spa_graph_node *node = &this->port->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->port->rt.mix_port.io;
@@ -631,10 +634,11 @@ static int schedule_mix_output(void *data)
 	return SPA_RESULT_NEED_BUFFER;
 }
 
-static const struct spa_graph_node_callbacks schedule_mix_node = {
+static const struct spa_node schedule_mix_node = {
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
+	SPA_VERSION_NODE,
-	schedule_mix_input,
+	NULL,
-	schedule_mix_output,
+	.process_input = schedule_mix_input,
+	.process_output = schedule_mix_output,
 };
 
 static void port_destroy(void *data)
@@ -771,6 +775,9 @@ pw_jack_node_add_port(struct pw_jack_node *node,
 
 	pw_port_add(port->port, node->node);
 
+	pd->mix_node = schedule_mix_node;
+
+
 	{
 		struct spa_buffer *b = &pd->buf;
 		struct type *t = &pd->node->type;
@@ -792,7 +799,7 @@ pw_jack_node_add_port(struct pw_jack_node *node,
 		port->port->state = PW_PORT_STATE_PAUSED;
 	}
 	if (direction == PW_DIRECTION_INPUT) {
-		spa_graph_node_set_callbacks(&port->port->rt.mix_node, &schedule_mix_node, port);
+		spa_graph_node_set_implementation(&port->port->rt.mix_node, &pd->mix_node);
 	}
 
 

src/pipewire/link.c

@@ -1132,8 +1132,8 @@ struct pw_link *pw_link_new(struct pw_core *core,
 			    0,
 			    &this->io);
 
-	this->rt.in_port.callbacks_data = this;
+	this->rt.in_port.scheduler_data = this;
-	this->rt.out_port.callbacks_data = this;
+	this->rt.out_port.scheduler_data = this;
 
 	/* nodes can be in different data loops so we do this twice */
 	pw_loop_invoke(output_node->data_loop, do_add_link,

src/pipewire/node.c

@@ -273,26 +273,6 @@ void pw_node_register(struct pw_node *this)
 	pw_node_update_state(this, PW_NODE_STATE_SUSPENDED, NULL);
 }
 
-static int
-graph_impl_process_input(void *data)
-{
-	struct pw_node *this = data;
-	return spa_node_process_input(this->node);
-}
-
-static int
-graph_impl_process_output(void *data)
-{
-	struct pw_node *this = data;
-	return spa_node_process_output(this->node);
-}
-
-static const struct spa_graph_node_callbacks graph_callbacks = {
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
-	.process_input = graph_impl_process_input,
-        .process_output = graph_impl_process_output,
-};
-
 struct pw_node *pw_node_new(struct pw_core *core,
 			    struct pw_resource *owner,
 			    struct pw_global *parent,
@@ -342,9 +322,6 @@ struct pw_node *pw_node_new(struct pw_core *core,
 	pw_map_init(&this->output_port_map, 64, 64);
 
 	spa_graph_node_init(&this->rt.node);
-	spa_graph_node_set_callbacks(&this->rt.node,
-				     &graph_callbacks,
-				     this);
 
 	return this;
 
@@ -446,9 +423,8 @@ static void node_reuse_buffer(void *data, uint32_t port_id, uint32_t buffer_id)
 		if (p->port_id != port_id)
 			continue;
 
-		pp = p->peer;
+		if ((pp = p->peer) != NULL)
-		if (pp && pp->callbacks->reuse_buffer)
+			spa_node_port_reuse_buffer(pp->node->implementation, pp->port_id, buffer_id);
-			pp->callbacks->reuse_buffer(pp->callbacks_data, buffer_id);
 		break;
 	}
 }
@@ -468,6 +444,7 @@ void pw_node_set_implementation(struct pw_node *node,
 {
 	node->node = spa_node;
 	spa_node_set_callbacks(node->node, &node_callbacks, node);
+	spa_graph_node_set_implementation(&node->rt.node, spa_node);
 }
 
 struct spa_node *pw_node_get_implementation(struct pw_node *node)

src/pipewire/port.c

@@ -28,6 +28,8 @@
 /** \cond */
 struct impl {
 	struct pw_port this;
+
+	struct spa_node mix_node;
 };
 /** \endcond */
 
@@ -41,9 +43,10 @@ static void port_update_state(struct pw_port *port, enum pw_port_state state)
 	}
 }
 
-static int schedule_tee_input(void *data)
+static int schedule_tee_input(struct spa_node *data)
 {
-        struct pw_port *this = data;
+	struct impl *impl = SPA_CONTAINER_OF(data, struct impl, mix_node);
+        struct pw_port *this = &impl->this;
 	struct spa_graph_node *node = &this->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->rt.mix_port.io;
@@ -63,9 +66,10 @@ static int schedule_tee_input(void *data)
 	}
         return res;
 }
-static int schedule_tee_output(void *data)
+static int schedule_tee_output(struct spa_node *data)
 {
-        struct pw_port *this = data;
+	struct impl *impl = SPA_CONTAINER_OF(data, struct impl, mix_node);
+        struct pw_port *this = &impl->this;
 	struct spa_graph_node *node = &this->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->rt.mix_port.io;
@@ -77,25 +81,23 @@ static int schedule_tee_output(void *data)
 	return SPA_RESULT_NEED_BUFFER;
 }
 
-static const struct spa_graph_node_callbacks schedule_tee_node = {
+static int schedule_tee_reuse_buffer(struct spa_node *data, uint32_t port_id, uint32_t buffer_id)
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
-	schedule_tee_input,
-	schedule_tee_output,
-};
-
-static int schedule_tee_reuse_buffer(void *data, uint32_t buffer_id)
 {
 	return SPA_RESULT_OK;
 }
 
-static const struct spa_graph_port_callbacks schedule_tee_port = {
+static const struct spa_node schedule_tee_node = {
-	SPA_VERSION_GRAPH_PORT_CALLBACKS,
+	SPA_VERSION_NODE,
-	schedule_tee_reuse_buffer,
+	NULL,
+	.process_input = schedule_tee_input,
+	.process_output = schedule_tee_output,
+	.port_reuse_buffer = schedule_tee_reuse_buffer,
 };
 
-static int schedule_mix_input(void *data)
+static int schedule_mix_input(struct spa_node *data)
 {
-        struct pw_port *this = data;
+	struct impl *impl = SPA_CONTAINER_OF(data, struct impl, mix_node);
+        struct pw_port *this = &impl->this;
 	struct spa_graph_node *node = &this->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->rt.mix_port.io;
@@ -111,9 +113,10 @@ static int schedule_mix_input(void *data)
 	return SPA_RESULT_HAVE_BUFFER;
 }
 
-static int schedule_mix_output(void *data)
+static int schedule_mix_output(struct spa_node *data)
 {
-        struct pw_port *this = data;
+	struct impl *impl = SPA_CONTAINER_OF(data, struct impl, mix_node);
+        struct pw_port *this = &impl->this;
 	struct spa_graph_node *node = &this->rt.mix_node;
 	struct spa_graph_port *p;
 	struct spa_port_io *io = this->rt.mix_port.io;
@@ -126,19 +129,17 @@ static int schedule_mix_output(void *data)
 	return SPA_RESULT_NEED_BUFFER;
 }
 
-static const struct spa_graph_node_callbacks schedule_mix_node = {
+static int schedule_mix_reuse_buffer(struct spa_node *data, uint32_t port_id, uint32_t buffer_id)
-	SPA_VERSION_GRAPH_NODE_CALLBACKS,
-	schedule_mix_input,
-	schedule_mix_output,
-};
-
-static int schedule_mix_reuse_buffer(void *data, uint32_t buffer_id)
 {
 	return SPA_RESULT_OK;
 }
-static const struct spa_graph_port_callbacks schedule_mix_port = {
+
-	SPA_VERSION_GRAPH_PORT_CALLBACKS,
+static const struct spa_node schedule_mix_node = {
-	schedule_mix_reuse_buffer,
+	SPA_VERSION_NODE,
+	NULL,
+	.process_input = schedule_mix_input,
+	.process_output = schedule_mix_output,
+	.port_reuse_buffer = schedule_mix_reuse_buffer,
 };
 
 struct pw_port *pw_port_new(enum pw_direction direction,
@@ -175,29 +176,24 @@ struct pw_port *pw_port_new(enum pw_direction direction,
 
 	spa_hook_list_init(&this->listener_list);
 
-	spa_graph_port_set_callbacks(&this->rt.port, NULL, this);
-
 	spa_graph_port_init(&this->rt.port,
 			    this->direction,
 			    this->port_id,
 			    0,
 			    &this->io);
 	spa_graph_node_init(&this->rt.mix_node);
-	spa_graph_node_set_callbacks(&this->rt.mix_node,
+
-				     this->direction == PW_DIRECTION_INPUT ?
+	impl->mix_node = this->direction == PW_DIRECTION_INPUT ?  schedule_mix_node : schedule_tee_node;
-					&schedule_mix_node :
+	spa_graph_node_set_implementation(&this->rt.mix_node, &impl->mix_node);
-					&schedule_tee_node,
-				     this);
 	spa_graph_port_init(&this->rt.mix_port,
 			    pw_direction_reverse(this->direction),
 			    0,
 			    0,
 			    &this->io);
-	spa_graph_port_set_callbacks(&this->rt.mix_port,
+
-				     this->direction == PW_DIRECTION_INPUT ?
+	this->rt.mix_port.scheduler_data = this;
-					&schedule_mix_port :
+	this->rt.port.scheduler_data = this;
-					&schedule_tee_port,
+
-				     this);
 	return this;
 
        no_mem:

src/pipewire/remote.c

@@ -78,6 +78,7 @@ struct node_data {
 	struct spa_hook node_listener;
 
         struct pw_client_node_proxy *node_proxy;
+	struct spa_hook node_proxy_listener;
 	struct spa_hook proxy_listener;
 
         struct pw_array mem_ids;
@@ -430,7 +431,7 @@ static void handle_rtnode_message(struct pw_proxy *proxy, struct pw_client_node_
 		/* process all input in the mixers */
 		spa_list_for_each(port, &n->ports[SPA_DIRECTION_INPUT], link) {
 			pn = port->peer->node;
-	                pn->state = pn->callbacks->process_input(pn->callbacks_data);
+	                pn->state = spa_node_process_input(pn->implementation);
 	                if (pn->state == SPA_RESULT_HAVE_BUFFER)
 	                        spa_graph_have_output(data->node->rt.graph, pn);
 			else {
@@ -444,7 +445,7 @@ static void handle_rtnode_message(struct pw_proxy *proxy, struct pw_client_node_
 		}
         }
 	else if (PW_CLIENT_NODE_MESSAGE_TYPE(message) == PW_CLIENT_NODE_MESSAGE_PROCESS_OUTPUT) {
-		n->callbacks->process_output(n->callbacks_data);
+		spa_node_process_output(n->implementation);
 	}
 	else if (PW_CLIENT_NODE_MESSAGE_TYPE(message) == PW_CLIENT_NODE_MESSAGE_REUSE_BUFFER) {
 	}
@@ -479,6 +480,28 @@ on_rtsocket_condition(void *user_data, int fd, enum spa_io mask)
 	}
 }
 
+static void clean_transport(struct pw_proxy *proxy)
+{
+	struct node_data *data = proxy->user_data;
+	struct pw_port *port;
+
+	if (data->trans == NULL)
+		return;
+
+	spa_list_for_each(port, &data->node->input_ports, link)
+		spa_graph_port_remove(&data->in_ports[port->port_id]);
+	spa_list_for_each(port, &data->node->output_ports, link)
+		spa_graph_port_remove(&data->out_ports[port->port_id]);
+
+	free(data->in_ports);
+	free(data->out_ports);
+	pw_client_node_transport_destroy(data->trans);
+	unhandle_socket(proxy);
+	close(data->rtwritefd);
+
+	data->trans = NULL;
+}
+
 static void client_node_transport(void *object, uint32_t node_id,
                                   int readfd, int writefd,
 				  struct pw_client_node_transport *transport)
@@ -488,14 +511,18 @@ static void client_node_transport(void *object, uint32_t node_id,
 	struct pw_port *port;
 	int i;
 
+	clean_transport(proxy);
+
 	data->node_id = node_id;
 	data->trans = transport;
 
 	pw_log_info("remote-node %p: create transport %p with fds %d %d for node %u",
 		proxy, data->trans, readfd, writefd, node_id);
 
-	data->in_ports = calloc(data->trans->area->max_input_ports, sizeof(struct spa_graph_port));
+	data->in_ports = calloc(data->trans->area->max_input_ports,
-	data->out_ports = calloc(data->trans->area->max_output_ports, sizeof(struct spa_graph_port));
+				 sizeof(struct spa_graph_port));
+	data->out_ports = calloc(data->trans->area->max_output_ports,
+				  sizeof(struct spa_graph_port));
 
 	for (i = 0; i < data->trans->area->max_input_ports; i++) {
 		spa_graph_port_init(&data->in_ports[i],
@@ -521,8 +548,6 @@ static void client_node_transport(void *object, uint32_t node_id,
 
         data->rtreadfd = readfd;
         data->rtwritefd = writefd;
-
-	unhandle_socket(proxy);
         data->rtsocket_source = pw_loop_add_io(proxy->remote->core->data_loop,
                                                data->rtreadfd,
                                                SPA_IO_ERR | SPA_IO_HUP,
@@ -981,6 +1006,25 @@ static const struct pw_node_events node_events = {
 	.have_output = node_have_output,
 };
 
+static void node_proxy_destroy(void *data)
+{
+	struct node_data *d = data;
+	struct pw_proxy *proxy = (struct pw_proxy*) d->node_proxy;
+
+	clean_transport(proxy);
+	clear_buffers(proxy);
+	clear_mems(proxy);
+	pw_array_clear(&d->mem_ids);
+	pw_array_clear(&d->buffer_ids);
+
+	spa_hook_remove(&d->node_listener);
+}
+
+static const struct pw_proxy_events proxy_events = {
+	PW_VERSION_PROXY_EVENTS,
+	.destroy = node_proxy_destroy,
+};
+
 struct pw_proxy *pw_remote_export(struct pw_remote *remote,
 				  struct pw_node *node)
 {
@@ -1010,10 +1054,11 @@ struct pw_proxy *pw_remote_export(struct pw_remote *remote,
         pw_array_init(&data->buffer_ids, 32);
         pw_array_ensure_size(&data->buffer_ids, sizeof(struct buffer_id) * 64);
 
+	pw_proxy_add_listener(proxy, &data->proxy_listener, &proxy_events, data);
 	pw_node_add_listener(node, &data->node_listener, &node_events, data);
 
         pw_client_node_proxy_add_listener(data->node_proxy,
-					  &data->proxy_listener,
+					  &data->node_proxy_listener,
 					  &client_node_events,
 					  proxy);
         do_node_init(proxy);