gstreamer/gst/tcp/gstmultifdsink.c
Thomas Vander Stichele 6dc005fbed change log level
Original commit message from CVS:
change log level
2004-06-27 12:22:32 +00:00

1042 lines
31 KiB
C

/* GStreamer
* Copyright (C) <1999> Erik Walthinsen <omega@cse.ogi.edu>
* Copyright (C) <2004> Thomas Vander Stichele <thomas at apestaart dot org>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library 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
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library 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.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <gst/gst-i18n-plugin.h>
#include <sys/ioctl.h>
#ifdef HAVE_FIONREAD_IN_SYS_FILIO
#include <sys/filio.h>
#endif
#include "gstmultifdsink.h"
#include "gsttcp-marshal.h"
#define CONTROL_RESTART 'R' /* restart the select call */
#define CONTROL_STOP 'S' /* stop the select call */
#define CONTROL_SOCKETS(sink) sink->control_sock
#define WRITE_SOCKET(sink) sink->control_sock[1]
#define READ_SOCKET(sink) sink->control_sock[0]
#define SEND_COMMAND(sink, command) \
G_STMT_START { \
unsigned char c; c = command; \
write (WRITE_SOCKET(sink), &c, 1); \
} G_STMT_END
#define READ_COMMAND(sink, command) \
G_STMT_START { \
read(READ_SOCKET(sink), &command, 1); \
} G_STMT_END
/* elementfactory information */
static GstElementDetails gst_multifdsink_details =
GST_ELEMENT_DETAILS ("TCP Server sink",
"Sink/Network",
"Send data as a server over the network via TCP",
"Thomas Vander Stichele <thomas at apestaart dot org>");
GST_DEBUG_CATEGORY (multifdsink_debug);
#define GST_CAT_DEFAULT (multifdsink_debug)
/* MultiFdSink signals and args */
enum
{
/* methods */
SIGNAL_ADD,
SIGNAL_REMOVE,
SIGNAL_CLEAR,
/* signals */
SIGNAL_CLIENT_ADDED,
SIGNAL_CLIENT_REMOVED,
LAST_SIGNAL
};
/* this is really arbitrary choosen */
#define DEFAULT_BUFFERS_MAX -1
#define DEFAULT_BUFFERS_SOFT_MAX -1
enum
{
ARG_0,
ARG_PROTOCOL,
ARG_BUFFERS_MAX,
ARG_BUFFERS_SOFT_MAX,
ARG_BUFFERS_QUEUED,
ARG_RECOVER_POLICY,
};
#define GST_TYPE_RECOVER_POLICY (gst_recover_policy_get_type())
static GType
gst_recover_policy_get_type (void)
{
static GType recover_policy_type = 0;
static GEnumValue recover_policy[] = {
{GST_RECOVER_POLICY_NONE, "GST_RECOVER_POLICY_NONE",
"Do not try to recover"},
{GST_RECOVER_POLICY_RESYNC_START, "GST_RECOVER_POLICY_RESYNC_START",
"Resync client to most recent buffer"},
{GST_RECOVER_POLICY_RESYNC_SOFT, "GST_RECOVER_POLICY_RESYNC_SOFT",
"Resync client to soft limit"},
{GST_RECOVER_POLICY_RESYNC_KEYFRAME, "GST_RECOVER_POLICY_RESYNC_KEYFRAME",
"Resync client to most recent keyframe"},
{0, NULL, NULL},
};
if (!recover_policy_type) {
recover_policy_type =
g_enum_register_static ("GstTCPRecoverPolicy", recover_policy);
}
return recover_policy_type;
}
static void gst_multifdsink_base_init (gpointer g_class);
static void gst_multifdsink_class_init (GstMultiFdSinkClass * klass);
static void gst_multifdsink_init (GstMultiFdSink * multifdsink);
static void gst_multifdsink_client_remove (GstMultiFdSink * sink,
GstTCPClient * client);
static void gst_multifdsink_chain (GstPad * pad, GstData * _data);
static GstElementStateReturn gst_multifdsink_change_state (GstElement *
element);
static void gst_multifdsink_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec);
static void gst_multifdsink_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec);
static GstElementClass *parent_class = NULL;
static guint gst_multifdsink_signals[LAST_SIGNAL] = { 0 };
GType
gst_multifdsink_get_type (void)
{
static GType multifdsink_type = 0;
if (!multifdsink_type) {
static const GTypeInfo multifdsink_info = {
sizeof (GstMultiFdSinkClass),
gst_multifdsink_base_init,
NULL,
(GClassInitFunc) gst_multifdsink_class_init,
NULL,
NULL,
sizeof (GstMultiFdSink),
0,
(GInstanceInitFunc) gst_multifdsink_init,
NULL
};
multifdsink_type =
g_type_register_static (GST_TYPE_ELEMENT, "GstMultiFdSink",
&multifdsink_info, 0);
}
return multifdsink_type;
}
static void
gst_multifdsink_base_init (gpointer g_class)
{
GstElementClass *element_class = GST_ELEMENT_CLASS (g_class);
gst_element_class_set_details (element_class, &gst_multifdsink_details);
}
static void
gst_multifdsink_class_init (GstMultiFdSinkClass * klass)
{
GObjectClass *gobject_class;
GstElementClass *gstelement_class;
gobject_class = (GObjectClass *) klass;
gstelement_class = (GstElementClass *) klass;
parent_class = g_type_class_ref (GST_TYPE_ELEMENT);
g_object_class_install_property (gobject_class, ARG_PROTOCOL,
g_param_spec_enum ("protocol", "Protocol", "The protocol to wrap data in",
GST_TYPE_TCP_PROTOCOL_TYPE, GST_TCP_PROTOCOL_TYPE_NONE,
G_PARAM_READWRITE));
g_object_class_install_property (G_OBJECT_CLASS (klass), ARG_BUFFERS_MAX,
g_param_spec_int ("buffers-max", "Buffers max",
"max number of buffers to queue (-1 = no limit)", -1, G_MAXINT,
DEFAULT_BUFFERS_MAX, G_PARAM_READWRITE));
g_object_class_install_property (G_OBJECT_CLASS (klass), ARG_BUFFERS_SOFT_MAX,
g_param_spec_int ("buffers-soft-max", "Buffers soft max",
"Recover client when going over this limit (-1 = no limit)", -1,
G_MAXINT, DEFAULT_BUFFERS_SOFT_MAX, G_PARAM_READWRITE));
g_object_class_install_property (G_OBJECT_CLASS (klass), ARG_BUFFERS_QUEUED,
g_param_spec_int ("buffers-queued", "Buffers queued",
"Number of buffers current queued", 0, G_MAXINT, 0,
G_PARAM_READABLE));
g_object_class_install_property (gobject_class, ARG_RECOVER_POLICY,
g_param_spec_enum ("recover-policy", "Recover Policy",
"How to recover when client reaches the soft max",
GST_TYPE_RECOVER_POLICY, GST_RECOVER_POLICY_NONE, G_PARAM_READWRITE));
gst_multifdsink_signals[SIGNAL_ADD] =
g_signal_new ("add", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstMultiFdSinkClass, add),
NULL, NULL, g_cclosure_marshal_VOID__INT, G_TYPE_NONE, 1, G_TYPE_INT);
gst_multifdsink_signals[SIGNAL_REMOVE] =
g_signal_new ("remove", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstMultiFdSinkClass, remove),
NULL, NULL, g_cclosure_marshal_VOID__INT, G_TYPE_NONE, 1, G_TYPE_INT);
gst_multifdsink_signals[SIGNAL_CLEAR] =
g_signal_new ("clear", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstMultiFdSinkClass, clear),
NULL, NULL, g_cclosure_marshal_VOID__INT, G_TYPE_NONE, 0);
gst_multifdsink_signals[SIGNAL_CLIENT_ADDED] =
g_signal_new ("client-added", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstMultiFdSinkClass, client_added),
NULL, NULL, gst_tcp_marshal_VOID__STRING_UINT, G_TYPE_NONE, 2,
G_TYPE_STRING, G_TYPE_UINT);
gst_multifdsink_signals[SIGNAL_CLIENT_REMOVED] =
g_signal_new ("client-removed", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstMultiFdSinkClass,
client_removed), NULL, NULL, gst_tcp_marshal_VOID__STRING_UINT,
G_TYPE_NONE, 2, G_TYPE_STRING, G_TYPE_UINT);
gobject_class->set_property = gst_multifdsink_set_property;
gobject_class->get_property = gst_multifdsink_get_property;
gstelement_class->change_state = gst_multifdsink_change_state;
klass->add = gst_multifdsink_add;
klass->remove = gst_multifdsink_remove;
klass->clear = gst_multifdsink_clear;
GST_DEBUG_CATEGORY_INIT (multifdsink_debug, "multifdsink", 0, "FD sink");
}
static void
gst_multifdsink_init (GstMultiFdSink * this)
{
/* create the sink pad */
this->sinkpad = gst_pad_new ("sink", GST_PAD_SINK);
gst_element_add_pad (GST_ELEMENT (this), this->sinkpad);
gst_pad_set_chain_function (this->sinkpad, gst_multifdsink_chain);
GST_FLAG_UNSET (this, GST_MULTIFDSINK_OPEN);
this->protocol = GST_TCP_PROTOCOL_TYPE_NONE;
this->clients = NULL;
this->bufqueue = g_array_new (FALSE, TRUE, sizeof (GstBuffer *));
this->buffers_max = DEFAULT_BUFFERS_MAX;
this->buffers_soft_max = DEFAULT_BUFFERS_SOFT_MAX;
this->clientslock = g_mutex_new ();
this->recover_policy = GST_RECOVER_POLICY_NONE;
}
static void
gst_multifdsink_debug_fdset (GstMultiFdSink * sink, fd_set * testfds)
{
int fd;
for (fd = 0; fd < FD_SETSIZE; fd++) {
if (FD_ISSET (fd, testfds)) {
GST_LOG_OBJECT (sink, "fd %d", fd);
}
}
}
void
gst_multifdsink_add (GstMultiFdSink * sink, int fd)
{
GstTCPClient *client;
/* create client datastructure */
client = g_new0 (GstTCPClient, 1);
client->fd = fd;
client->bufpos = -1;
client->bufoffset = 0;
client->sending = NULL;
g_mutex_lock (sink->clientslock);
sink->clients = g_list_prepend (sink->clients, client);
g_mutex_unlock (sink->clientslock);
/* we always read from a client */
FD_SET (fd, &sink->readfds);
/* set the socket to non blocking */
fcntl (fd, F_SETFL, O_NONBLOCK);
g_signal_emit (G_OBJECT (sink),
gst_multifdsink_signals[SIGNAL_CLIENT_ADDED], 0, NULL, fd);
}
void
gst_multifdsink_remove (GstMultiFdSink * sink, int fd)
{
GList *clients;
g_mutex_lock (sink->clientslock);
/* loop over the clients to find the one with the fd */
for (clients = sink->clients; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
client = (GstTCPClient *) clients->data;
if (client->fd == fd) {
gst_multifdsink_client_remove (sink, client);
break;
}
}
g_mutex_unlock (sink->clientslock);
}
void
gst_multifdsink_clear (GstMultiFdSink * sink)
{
GList *clients;
g_mutex_lock (sink->clientslock);
for (clients = sink->clients; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
client = (GstTCPClient *) clients->data;
gst_multifdsink_client_remove (sink, client);
}
g_mutex_unlock (sink->clientslock);
}
static void
gst_multifdsink_client_remove (GstMultiFdSink * sink, GstTCPClient * client)
{
int fd = client->fd;
/* FIXME: if we keep track of ip we can log it here and signal */
GST_DEBUG_OBJECT (sink, "removing client on fd %d", fd);
if (close (fd) != 0) {
GST_DEBUG_OBJECT (sink, "error closing fd %d: %s", fd, g_strerror (errno));
}
FD_CLR (fd, &sink->readfds);
FD_CLR (fd, &sink->writefds);
sink->clients = g_list_remove (sink->clients, client);
g_free (client);
g_signal_emit (G_OBJECT (sink),
gst_multifdsink_signals[SIGNAL_CLIENT_REMOVED], 0, NULL, fd);
}
/* handle a read on a client fd,
* which either indicates a close or should be ignored
* returns FALSE if the client has been closed. */
static gboolean
gst_multifdsink_handle_client_read (GstMultiFdSink * sink,
GstTCPClient * client)
{
int nread, fd;
fd = client->fd;
GST_LOG_OBJECT (sink, "select reports client read on fd %d", fd);
ioctl (fd, FIONREAD, &nread);
if (nread == 0) {
/* client sent close, so remove it */
GST_DEBUG_OBJECT (sink, "client asked for close, removing on fd %d", fd);
return FALSE;
} else {
/* FIXME: we should probably just Read 'n' Drop */
g_warning ("Don't know what to do with %d bytes to read", nread);
}
return TRUE;
}
static gboolean
gst_multifdsink_client_queue_data (GstMultiFdSink * sink, GstTCPClient * client,
gchar * data, gint len)
{
GstBuffer *buf;
buf = gst_buffer_new ();
GST_BUFFER_DATA (buf) = data;
GST_BUFFER_SIZE (buf) = len;
GST_LOG_OBJECT (sink, "Queueing data of length %d for fd %d",
len, client->fd);
client->sending = g_list_append (client->sending, buf);
return TRUE;
}
static gboolean
gst_multifdsink_client_queue_caps (GstMultiFdSink * sink, GstTCPClient * client,
const GstCaps * caps)
{
guint8 *header;
guint8 *payload;
guint length;
gchar *string;
string = gst_caps_to_string (caps);
GST_DEBUG_OBJECT (sink, "Queueing caps %s for fd %d through GDP", string,
client->fd);
g_free (string);
if (!gst_dp_packet_from_caps (caps, 0, &length, &header, &payload)) {
GST_DEBUG_OBJECT (sink, "Could not create GDP packet from caps");
return FALSE;
}
gst_multifdsink_client_queue_data (sink, client, header, length);
length = gst_dp_header_payload_length (header);
gst_multifdsink_client_queue_data (sink, client, payload, length);
return TRUE;
}
static gboolean
gst_multifdsink_client_queue_buffer (GstMultiFdSink * sink,
GstTCPClient * client, GstBuffer * buffer)
{
if (sink->protocol == GST_TCP_PROTOCOL_TYPE_GDP) {
guint8 *header;
guint len;
if (!gst_dp_header_from_buffer (buffer, 0, &len, &header)) {
GST_DEBUG_OBJECT (sink,
"could not create header, removing client on fd %d", client->fd);
return FALSE;
}
gst_multifdsink_client_queue_data (sink, client, header, len);
}
gst_buffer_ref (buffer);
client->sending = g_list_append (client->sending, buffer);
return TRUE;
}
/* handle a write on a client,
* which indicates a read request from a client.
*
* The strategy is as follows, for each client we maintain a queue of GstBuffers
* that contain the raw bytes we need to send to the client. In the case of the
* GDP protocol, we create buffers out of the header bytes so that we can only focus
* on sending buffers.
*
* We first check to see if we need to send caps (in GDP) and streamheaders. If so,
* we queue them.
*
* Then we run into the main loop that tries to send as many buffers as possible. It
* will first exhaust the client->sending queue and if the queue is empty, it will
* pick a buffer from the global queue.
*
* Sending the Buffers from the client->sending queue is basically writing the bytes
* to the socket and maintaining a count of the bytes that were sent. When the buffer
* is completely sent, it is removed from the client->sending queue and we try to pick
* a new buffer for sending.
*
* When the sending returns a partial buffer we stop sending more data as the next send
* operation could block.
*/
static gboolean
gst_multifdsink_handle_client_write (GstMultiFdSink * sink,
GstTCPClient * client)
{
int fd = client->fd;
gboolean more;
gboolean res;
/* when using GDP, first check if we have queued caps yet */
if (sink->protocol == GST_TCP_PROTOCOL_TYPE_GDP) {
if (!client->caps_sent) {
const GstCaps *caps = GST_PAD_CAPS (GST_PAD_PEER (sink->sinkpad));
/* queue caps for sending */
res = gst_multifdsink_client_queue_caps (sink, client, caps);
if (!res) {
GST_DEBUG_OBJECT (sink, "Failed queueing caps, removing client");
return FALSE;
}
client->caps_sent = TRUE;
}
}
/* if we have streamheader buffers, and haven't sent them to this client
* yet, send them out one by one */
if (!client->streamheader_sent) {
if (sink->streamheader) {
GList *l;
for (l = sink->streamheader; l; l = l->next) {
/* queue stream headers for sending */
res =
gst_multifdsink_client_queue_buffer (sink, client,
GST_BUFFER (l->data));
if (!res) {
GST_DEBUG_OBJECT (sink,
"Failed queueing streamheader, removing client");
return FALSE;
}
}
}
client->streamheader_sent = TRUE;
}
more = TRUE;
do {
gint maxsize;
if (!client->sending) {
/* client is not working on a buffer */
if (client->bufpos == -1) {
/* client is too fast, remove from write queue until new buffer is
* available */
FD_CLR (fd, &sink->writefds);
return TRUE;
} else {
/* client can pick a buffer from the global queue */
GstBuffer *buf;
/* grab buffer and ref, we need to ref since it could be unreffed in
* another thread */
buf = g_array_index (sink->bufqueue, GstBuffer *, client->bufpos);
client->bufpos--;
/* queueing a buffer will ref it */
gst_multifdsink_client_queue_buffer (sink, client, buf);
/* need to start from the first byte for this new buffer */
client->bufoffset = 0;
}
}
/* see if we need to send something */
if (client->sending) {
ssize_t wrote;
GstBuffer *head;
/* pick first buffer from list */
head = GST_BUFFER (client->sending->data);
maxsize = GST_BUFFER_SIZE (head) - client->bufoffset;
/* try to write the complete buffer */
wrote =
send (fd, GST_BUFFER_DATA (head) + client->bufoffset, maxsize,
MSG_NOSIGNAL);
if (wrote < 0) {
/* hmm error.. */
if (errno == EAGAIN) {
/* nothing serious, resource was unavailable, try again later */
more = FALSE;
} else {
GST_DEBUG_OBJECT (sink, "could not write, removing client on fd %d",
fd);
return FALSE;
}
} else if (wrote < maxsize) {
/* partial write means that the client cannot read more and we should
* stop sending more */
GST_LOG_OBJECT (sink, "partial write on %d of %d bytes", fd, wrote);
client->bufoffset += wrote;
more = FALSE;
} else {
/* complete buffer was written, we can proceed to the next one */
client->sending = g_list_remove (client->sending, head);
gst_buffer_unref (head);
/* make sure we start from byte 0 for the next buffer */
client->bufoffset = 0;
}
}
} while (more);
return TRUE;
}
/* calculate the new position for a client after recovery. This function
* does not update the client position but merely returns the required
* position.
*/
static gint
gst_multifdsink_recover_client (GstMultiFdSink * sink, GstTCPClient * client)
{
gint newbufpos;
/* FIXME: implement recover procedure here, like moving the position to
* the next keyframe, dropping buffers back to the beginning of the queue,
* stuff like that... */
GST_WARNING_OBJECT (sink, "client %p with fd %d is lagging",
client, client->fd);
switch (sink->recover_policy) {
case GST_RECOVER_POLICY_NONE:
/* do nothing, client will catch up or get kicked out when it reaches
* the hard max */
newbufpos = client->bufpos;
break;
case GST_RECOVER_POLICY_RESYNC_START:
/* move to beginning of queue */
newbufpos = -1;
break;
case GST_RECOVER_POLICY_RESYNC_SOFT:
/* move to beginning of soft max */
newbufpos = sink->buffers_soft_max;
break;
case GST_RECOVER_POLICY_RESYNC_KEYFRAME:
/* FIXME, find keyframe in buffers */
newbufpos = sink->buffers_soft_max;
break;
default:
newbufpos = sink->buffers_soft_max;
break;
}
return newbufpos;
}
/* Queue a buffer on the global queue.
*
* This functions adds the buffer to the front of a GArray. It removes the
* tail buffer if the max queue size is exceeded. Unreffing the buffer that
* is queued. Note that unreffing the buffer is not a problem as clients who
* started writing out this buffer will still have a reference to it in the
* client->sending queue.
*
* After adding the buffer, we update all client positions in the queue. If
* a client moves of the soft max, we start the recovery procedure for this
* slow client. If it goes over the hard max, it is put into the slow list
* and removed.
*
* Special care is taken of clients that were waiting for a new buffer (they
* had a position of -1) because they can proceed after adding this new buffer.
* This is done by adding the client back into the write fd_set and signalling
* the select thread that the fd_set changed.
*
*/
static void
gst_multifdsink_queue_buffer (GstMultiFdSink * sink, GstBuffer * buf)
{
GList *clients;
gint queuelen;
GList *slow = NULL;
gboolean need_signal = FALSE;
gint max_buffer_usage;
gint i;
g_mutex_lock (sink->clientslock);
/* add buffer to queue */
g_array_prepend_val (sink->bufqueue, buf);
queuelen = sink->bufqueue->len;
/* then loop over the clients and update the positions */
max_buffer_usage = 0;
for (clients = sink->clients; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
client = (GstTCPClient *) clients->data;
client->bufpos++;
GST_LOG_OBJECT (sink, "client %p with fd %d at position %d",
client, client->fd, client->bufpos);
/* check soft max if needed, recover client */
if (sink->buffers_soft_max > 0 && client->bufpos >= sink->buffers_soft_max) {
gint newpos;
newpos = gst_multifdsink_recover_client (sink, client);
if (newpos != client->bufpos) {
client->bufpos = newpos;
client->discont = TRUE;
}
}
/* check hard max, remove client */
if (sink->buffers_max > 0 && client->bufpos >= sink->buffers_max) {
/* remove client */
GST_WARNING_OBJECT (sink, "client %p with fd %d is too slow, removing",
client, client->fd);
FD_CLR (client->fd, &sink->readfds);
FD_CLR (client->fd, &sink->writefds);
slow = g_list_prepend (slow, client);
/* cannot send data to this client anymore. need to signal the select thread that
* the fd_set changed */
need_signal = TRUE;
/* set client to invalid position while being removed */
client->bufpos = -1;
} else if (client->bufpos == 0) {
/* can send data to this client now. need to signal the select thread that
* the fd_set changed */
FD_SET (client->fd, &sink->writefds);
need_signal = TRUE;
}
/* keep track of maximum buffer usage */
if (client->bufpos > max_buffer_usage) {
max_buffer_usage = client->bufpos;
}
}
/* remove crap clients */
for (clients = slow; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
client = (GstTCPClient *) slow->data;
gst_multifdsink_client_remove (sink, client);
}
g_list_free (slow);
/* nobody is referencing buffers after max_buffer_usage so we can
* remove them from the queue */
for (i = queuelen - 1; i > max_buffer_usage; i--) {
GstBuffer *old;
/* queue exceeded max size */
queuelen--;
old = g_array_index (sink->bufqueue, GstBuffer *, i);
sink->bufqueue = g_array_remove_index (sink->bufqueue, i);
/* unref tail buffer */
gst_buffer_unref (old);
}
sink->buffers_queued = max_buffer_usage;
g_mutex_unlock (sink->clientslock);
/* and send a signal to thread if fd_set changed */
if (need_signal) {
SEND_COMMAND (sink, CONTROL_RESTART);
}
}
/* Handle the clients. Basically does a blocking select for one
* of the client fds to become read or writable. We also have a
* filedescriptor to receive commands on that we need to check.
*
* After going out of the select call, we read and write to all
* clients that can do so. Badly behaving clients are put on a
* garbage list and removed.
*/
static void
gst_multifdsink_handle_clients (GstMultiFdSink * sink)
{
int result;
fd_set testreadfds, testwritefds;
GList *clients, *error = NULL;
gboolean try_again;
GstMultiFdSinkClass *fclass;
fclass = GST_MULTIFDSINK_GET_CLASS (sink);
do {
try_again = FALSE;
/* check for:
* - server socket input (ie, new client connections)
* - client socket input (ie, clients saying goodbye)
* - client socket output (ie, client reads) */
testwritefds = sink->writefds;
testreadfds = sink->readfds;
GST_LOG_OBJECT (sink, "doing select on server + client fds for reads");
gst_multifdsink_debug_fdset (sink, &testreadfds);
GST_LOG_OBJECT (sink, "doing select on client fds for writes");
gst_multifdsink_debug_fdset (sink, &testwritefds);
result =
select (FD_SETSIZE, &testreadfds, &testwritefds, (fd_set *) 0, NULL);
/* < 0 is an error, 0 just means a timeout happened */
if (result < 0) {
GST_ELEMENT_ERROR (sink, RESOURCE, READ, (NULL),
("select failed: %s", g_strerror (errno)));
return;
}
GST_LOG_OBJECT (sink, "%d sockets had action", result);
GST_LOG_OBJECT (sink, "done select on server/client fds for reads");
gst_multifdsink_debug_fdset (sink, &testreadfds);
GST_LOG_OBJECT (sink, "done select on client fds for writes");
gst_multifdsink_debug_fdset (sink, &testwritefds);
if (FD_ISSET (READ_SOCKET (sink), &testreadfds)) {
gchar command;
READ_COMMAND (sink, command);
switch (command) {
case CONTROL_RESTART:
/* need to restart the select call as the fd_set changed */
try_again = TRUE;
break;
case CONTROL_STOP:
/* stop this function */
return;
default:
g_warning ("multifdsink: unknown control message received");
break;
}
}
} while (try_again);
if (fclass->select)
fclass->select (sink, &testreadfds, &testwritefds);
/* Check the reads */
g_mutex_lock (sink->clientslock);
for (clients = sink->clients; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
int fd;
client = (GstTCPClient *) clients->data;
fd = client->fd;
if (FD_ISSET (fd, &testreadfds)) {
/* handle client read */
if (!gst_multifdsink_handle_client_read (sink, client)) {
error = g_list_prepend (error, client);
continue;
}
}
if (FD_ISSET (fd, &testwritefds)) {
/* handle client write */
if (!gst_multifdsink_handle_client_write (sink, client)) {
error = g_list_prepend (error, client);
continue;
}
}
}
/* remove crappy clients */
for (clients = error; clients; clients = g_list_next (clients)) {
GstTCPClient *client;
client = (GstTCPClient *) error->data;
GST_LOG_OBJECT (sink, "removing client %p with fd %d with errors", client,
client->fd);
gst_multifdsink_client_remove (sink, client);
}
g_list_free (error);
g_mutex_unlock (sink->clientslock);
}
static gpointer
gst_multifdsink_thread (GstMultiFdSink * sink)
{
while (sink->running) {
gst_multifdsink_handle_clients (sink);
}
return NULL;
}
static void
gst_multifdsink_chain (GstPad * pad, GstData * _data)
{
GstBuffer *buf = GST_BUFFER (_data);
GstMultiFdSink *sink;
g_return_if_fail (pad != NULL);
g_return_if_fail (GST_IS_PAD (pad));
g_return_if_fail (buf != NULL);
sink = GST_MULTIFDSINK (GST_OBJECT_PARENT (pad));
g_return_if_fail (GST_FLAG_IS_SET (sink, GST_MULTIFDSINK_OPEN));
if (GST_IS_EVENT (buf)) {
g_warning ("FIXME: handle events");
return;
}
/* if the incoming buffer is marked as IN CAPS, then we assume for now
* it's a streamheader that needs to be sent to each new client, so we
* put it on our internal list of streamheader buffers.
* After that we return, since we only send these out when we get
* non IN_CAPS buffers so we properly keep track of clients that got
* streamheaders. */
if (GST_BUFFER_FLAG_IS_SET (buf, GST_BUFFER_IN_CAPS)) {
GST_DEBUG_OBJECT (sink,
"appending IN_CAPS buffer with length %d to streamheader",
GST_BUFFER_SIZE (buf));
sink->streamheader = g_list_append (sink->streamheader, buf);
return;
}
/* queue the buffer */
gst_multifdsink_queue_buffer (sink, buf);
sink->data_written += GST_BUFFER_SIZE (buf);
}
static void
gst_multifdsink_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec)
{
GstMultiFdSink *multifdsink;
g_return_if_fail (GST_IS_MULTIFDSINK (object));
multifdsink = GST_MULTIFDSINK (object);
switch (prop_id) {
case ARG_PROTOCOL:
multifdsink->protocol = g_value_get_enum (value);
break;
case ARG_BUFFERS_MAX:
multifdsink->buffers_max = g_value_get_int (value);
break;
case ARG_BUFFERS_SOFT_MAX:
multifdsink->buffers_soft_max = g_value_get_int (value);
break;
case ARG_RECOVER_POLICY:
multifdsink->recover_policy = g_value_get_enum (value);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
gst_multifdsink_get_property (GObject * object, guint prop_id, GValue * value,
GParamSpec * pspec)
{
GstMultiFdSink *multifdsink;
g_return_if_fail (GST_IS_MULTIFDSINK (object));
multifdsink = GST_MULTIFDSINK (object);
switch (prop_id) {
case ARG_PROTOCOL:
g_value_set_enum (value, multifdsink->protocol);
break;
case ARG_BUFFERS_MAX:
g_value_set_int (value, multifdsink->buffers_max);
break;
case ARG_BUFFERS_SOFT_MAX:
g_value_set_int (value, multifdsink->buffers_soft_max);
break;
case ARG_BUFFERS_QUEUED:
g_value_set_int (value, multifdsink->buffers_queued);
break;
case ARG_RECOVER_POLICY:
g_value_set_enum (value, multifdsink->recover_policy);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
/* create a socket for sending to remote machine */
static gboolean
gst_multifdsink_init_send (GstMultiFdSink * this)
{
GstMultiFdSinkClass *fclass;
fclass = GST_MULTIFDSINK_GET_CLASS (this);
FD_ZERO (&this->readfds);
FD_ZERO (&this->writefds);
if (socketpair (PF_UNIX, SOCK_STREAM, 0, CONTROL_SOCKETS (this)) < 0) {
perror ("creating socket pair");
}
FD_SET (READ_SOCKET (this), &this->readfds);
this->streamheader = NULL;
this->data_written = 0;
if (fclass->init) {
fclass->init (this);
}
this->running = TRUE;
this->thread = g_thread_create ((GThreadFunc) gst_multifdsink_thread,
this, TRUE, NULL);
return TRUE;
}
static void
gst_multifdsink_close (GstMultiFdSink * this)
{
GstMultiFdSinkClass *fclass;
fclass = GST_MULTIFDSINK_GET_CLASS (this);
this->running = FALSE;
SEND_COMMAND (this, CONTROL_STOP);
g_thread_join (this->thread);
close (READ_SOCKET (this));
close (WRITE_SOCKET (this));
if (this->streamheader) {
GList *l;
for (l = this->streamheader; l; l = l->next) {
gst_buffer_unref (l->data);
}
g_list_free (this->streamheader);
}
if (fclass->close)
fclass->close (this);
}
static GstElementStateReturn
gst_multifdsink_change_state (GstElement * element)
{
GstMultiFdSink *sink;
g_return_val_if_fail (GST_IS_MULTIFDSINK (element), GST_STATE_FAILURE);
sink = GST_MULTIFDSINK (element);
switch (GST_STATE_TRANSITION (element)) {
case GST_STATE_NULL_TO_READY:
if (!GST_FLAG_IS_SET (element, GST_MULTIFDSINK_OPEN)) {
if (!gst_multifdsink_init_send (sink))
return GST_STATE_FAILURE;
GST_FLAG_SET (sink, GST_MULTIFDSINK_OPEN);
}
break;
case GST_STATE_READY_TO_PAUSED:
break;
case GST_STATE_PAUSED_TO_PLAYING:
break;
case GST_STATE_PLAYING_TO_PAUSED:
break;
case GST_STATE_PAUSED_TO_READY:
break;
case GST_STATE_READY_TO_NULL:
if (GST_FLAG_IS_SET (element, GST_MULTIFDSINK_OPEN)) {
gst_multifdsink_close (GST_MULTIFDSINK (element));
GST_FLAG_UNSET (sink, GST_MULTIFDSINK_OPEN);
}
break;
}
if (GST_ELEMENT_CLASS (parent_class)->change_state)
return GST_ELEMENT_CLASS (parent_class)->change_state (element);
return GST_STATE_SUCCESS;
}