mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-18 07:47:17 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/subprojects/gst-rtsp-server/gst/rtsp-server/rtsp-client.c

5427 lines
156 KiB
C
Raw Normal View History

Initial import
2008-10-09 12:29:12 +00:00
/* GStreamer
* Copyright (C) 2008 Wim Taymans <wim.taymans at gmail.com>
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
* Copyright (C) 2015 Centricular Ltd
* Author: Sebastian Dröge <sebastian@centricular.com>
Initial import
2008-10-09 12:29:12 +00:00
*
* 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
Fix FSF address
2012-11-04 00:14:25 +00:00
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301, USA.
Initial import
2008-10-09 12:29:12 +00:00
*/
docs: improve docs
2013-07-11 14:57:14 +00:00
/**
* SECTION:rtsp-client
* @short_description: A client connection state
* @see_also: #GstRTSPServer, #GstRTSPThreadPool
*
* The client object handles the connection with a client for as long as a TCP
* connection is open.
*
* A #GstRTSPClient is created by #GstRTSPServer when a new connection is
* accepted and it inherits the #GstRTSPMountPoints, #GstRTSPSessionPool,
* #GstRTSPAuth and #GstRTSPThreadPool from the server.
*
* The client connection should be configured with the #GstRTSPConnection using
* gst_rtsp_client_set_connection() before it can be attached to a #GMainContext
* using gst_rtsp_client_attach(). From then on the client will handle requests
* on the connection.
*
* Use gst_rtsp_client_session_filter() to iterate or modify all the
* #GstRTSPSession objects managed by the client object.
*
* Last reviewed on 2013-07-11 (1.0.0)
*/
libs: fix API export/import and 'inconsistent linkage' on MSVC Export rtsp-server library API in headers when we're building the library itself, otherwise import the API from the headers. This fixes linker warnings on Windows when building with MSVC. Fix up some missing config.h includes when building the lib which is needed to get the export api define from config.h https://bugzilla.gnome.org/show_bug.cgi?id=797185
2018-09-24 08:36:21 +00:00
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
Initial import
2008-10-09 12:29:12 +00:00
rtsp: move network includes where they are needed
2011-01-11 21:41:12 +00:00
#include <stdio.h>
#include <string.h>
Initial import
2008-10-09 12:29:12 +00:00
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
#include <gst/sdp/gstmikey.h>
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
#include <gst/rtsp/gstrtsp-enumtypes.h>
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
Initial import
2008-10-09 12:29:12 +00:00
#include "rtsp-client.h"
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
#include "rtsp-sdp.h"
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
#include "rtsp-params.h"
stream: refactor TCP backpressure handling The previous implementation stopped sending TCP messages to all clients when a single one stopped consuming them, which obviously created problems for shared media. Instead, we now manage a backlog in stream-transport, and slow clients are removed once this backlog exceeds a maximum duration, currently hardcoded. Fixes #80
2019-10-15 17:08:32 +00:00
#include "rtsp-server-internal.h"
Initial import
2008-10-09 12:29:12 +00:00
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
typedef enum
{
TUNNEL_STATE_UNKNOWN,
TUNNEL_STATE_GET,
TUNNEL_STATE_POST
} GstRTSPTunnelState;
Document locking and its order
2013-02-18 19:59:58 +00:00
/* locking order:
* send_lock, lock, tunnels_lock
*/
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
struct _GstRTSPClientPrivate
{
Document locking and its order
2013-02-18 19:59:58 +00:00
GMutex lock; /* protects everything else */
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
GMutex send_lock;
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
GMutex watch_lock;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPConnection *connection;
GstRTSPWatch *watch;
client: ref the context until rtsp watch is alive Fixes https://bugzilla.gnome.org/show_bug.cgi?id=731569
2014-06-12 11:49:17 +00:00
GMainContext *watch_context;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
gchar *server_ip;
gboolean is_ipv6;
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
/* protected by send_lock */
GstRTSPClientSendFunc send_func;
gpointer send_data;
GDestroyNotify send_notify;
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
GstRTSPClientSendMessagesFunc send_messages_func;
gpointer send_messages_data;
GDestroyNotify send_messages_notify;
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
GArray *data_seqs;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPSessionPool *session_pool;
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
gulong session_removed_id;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPMountPoints *mount_points;
GstRTSPAuth *auth;
client: add method to configure thread pool
2013-07-10 15:06:36 +00:00
GstRTSPThreadPool *thread_pool;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
/* used to cache the media in the last requested DESCRIBE so that
* we can pick it up in the next SETUP immediately */
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
gchar *path;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPMedia *media;
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
GHashTable *transports;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GList *sessions;
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
guint sessions_cookie;
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
gboolean drop_backlog;
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
gint post_session_timeout;
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
guint content_length_limit;
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
gboolean had_session;
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
GSource *rtsp_ctrl_timeout;
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
guint rtsp_ctrl_timeout_cnt;
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
/* The version currently being used */
GstRTSPVersion version;
GHashTable *pipelined_requests; /* pipelined_request_id -> session_id */
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
GstRTSPTunnelState tstate;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
};
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
typedef struct
{
guint8 channel;
guint seq;
} DataSeq;
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
static GMutex tunnels_lock;
Document locking and its order
2013-02-18 19:59:58 +00:00
static GHashTable *tunnels; /* protected by tunnels_lock */
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: make define for the WATCH_BACKLOG See https://bugzilla.gnome.org/show_bug.cgi?id=736322
2014-09-16 09:29:38 +00:00
#define WATCH_BACKLOG_SIZE 100
client: Use client transport settings for multicast if allowed. This patch makes it possible for the client to send transport settings for multicast (destination && ttl). Client settings must be explicitly allowed or the server will use its own settings. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=685561
2012-10-10 09:06:02 +00:00
#define DEFAULT_SESSION_POOL NULL
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
#define DEFAULT_MOUNT_POINTS NULL
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
#define DEFAULT_DROP_BACKLOG TRUE
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
#define DEFAULT_POST_SESSION_TIMEOUT -1
client: Use client transport settings for multicast if allowed. This patch makes it possible for the client to send transport settings for multicast (destination && ttl). Client settings must be explicitly allowed or the server will use its own settings. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=685561
2012-10-10 09:06:02 +00:00
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
#define RTSP_CTRL_CB_INTERVAL 1
#define RTSP_CTRL_TIMEOUT_VALUE 60
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
enum
{
PROP_0,
PROP_SESSION_POOL,
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
PROP_MOUNT_POINTS,
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
PROP_DROP_BACKLOG,
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
PROP_POST_SESSION_TIMEOUT,
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
PROP_LAST
};
client: emit signal when closing
2011-01-12 14:35:51 +00:00
enum
{
SIGNAL_CLOSED,
add new-session signal to rtsp-client (fixes #683058)
2012-08-30 19:03:27 +00:00
SIGNAL_NEW_SESSION,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_OPTIONS_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_OPTIONS_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_DESCRIBE_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_DESCRIBE_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_SETUP_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_SETUP_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_PLAY_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_PLAY_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_PAUSE_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_PAUSE_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_TEARDOWN_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_TEARDOWN_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_SET_PARAMETER_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_SET_PARAMETER_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_GET_PARAMETER_REQUEST,
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
SIGNAL_GET_PARAMETER_REQUEST,
Add handle-response signal for when we receive a GET_PARAMETER response
2013-08-08 14:57:42 +00:00
SIGNAL_HANDLE_RESPONSE,
client: emit a signal before sending a message Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728970
2014-05-01 04:17:06 +00:00
SIGNAL_SEND_MESSAGE,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_ANNOUNCE_REQUEST,
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
SIGNAL_ANNOUNCE_REQUEST,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
SIGNAL_PRE_RECORD_REQUEST,
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
SIGNAL_RECORD_REQUEST,
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
SIGNAL_CHECK_REQUIREMENTS,
client: emit signal when closing
2011-01-12 14:35:51 +00:00
SIGNAL_LAST
};
Use GStreamer's debugging subsystem
2009-11-21 18:20:23 +00:00
GST_DEBUG_CATEGORY_STATIC (rtsp_client_debug);
#define GST_CAT_DEFAULT rtsp_client_debug
client: emit signal when closing
2011-01-12 14:35:51 +00:00
static guint gst_rtsp_client_signals[SIGNAL_LAST] = { 0 };
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
static void gst_rtsp_client_get_property (GObject * object, guint propid,
GValue * value, GParamSpec * pspec);
static void gst_rtsp_client_set_property (GObject * object, guint propid,
const GValue * value, GParamSpec * pspec);
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
static void gst_rtsp_client_finalize (GObject * obj);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
static void rtsp_ctrl_timeout_remove (GstRTSPClient * client);
client: Fix crash in close handler The close handler could trigger a crash because it invalidated the watch_context while still leaving a source attached to it which would be cleaned up at a later point.
2018-12-04 13:12:04 +00:00
client: use more GIO Fixes https://bugzilla.gnome.org/show_bug.cgi?id=681593
2012-09-07 15:14:30 +00:00
static GstSDPMessage *create_sdp (GstRTSPClient * client, GstRTSPMedia * media);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
static gboolean handle_sdp (GstRTSPClient * client, GstRTSPContext * ctx,
GstRTSPMedia * media, GstSDPMessage * sdp);
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
static gboolean default_configure_client_media (GstRTSPClient * client,
GstRTSPMedia * media, GstRTSPStream * stream, GstRTSPContext * ctx);
rtsp-client: Make configure_client_transport virtual This patch makes configure_client_transport virtual. The functionality is needed to handle some weird clients sending multicast transport settings as url options. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702173
2013-06-17 14:18:37 +00:00
static gboolean default_configure_client_transport (GstRTSPClient * client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext * ctx, GstRTSPTransport * ct);
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
static GstRTSPResult default_params_set (GstRTSPClient * client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext * ctx);
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
static GstRTSPResult default_params_get (GstRTSPClient * client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext * ctx);
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
static gchar *default_make_path_from_uri (GstRTSPClient * client,
const GstRTSPUrl * uri);
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
static void client_session_removed (GstRTSPSessionPool * pool,
GstRTSPSession * session, GstRTSPClient * client);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
static GstRTSPStatusCode default_pre_signal_handler (GstRTSPClient * client,
GstRTSPContext * ctx);
static gboolean pre_signal_accumulator (GSignalInvocationHint * ihint,
GValue * return_accu, const GValue * handler_return, gpointer data);
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
Update for g_type_class_add_private() deprecation in recent GLib
2018-06-24 10:44:26 +00:00
G_DEFINE_TYPE_WITH_PRIVATE (GstRTSPClient, gst_rtsp_client, G_TYPE_OBJECT);
Initial import
2008-10-09 12:29:12 +00:00
static void
gst_rtsp_client_class_init (GstRTSPClientClass * klass)
{
GObjectClass *gobject_class;
gobject_class = G_OBJECT_CLASS (klass);
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
gobject_class->get_property = gst_rtsp_client_get_property;
gobject_class->set_property = gst_rtsp_client_set_property;
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
gobject_class->finalize = gst_rtsp_client_finalize;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
rtsp-client: make create_sdp virtual method Fixes https://bugzilla.gnome.org/show_bug.cgi?id=680173
2012-07-18 13:54:49 +00:00
klass->create_sdp = create_sdp;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
klass->handle_sdp = handle_sdp;
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
klass->configure_client_media = default_configure_client_media;
rtsp-client: Make configure_client_transport virtual This patch makes configure_client_transport virtual. The functionality is needed to handle some weird clients sending multicast transport settings as url options. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702173
2013-06-17 14:18:37 +00:00
klass->configure_client_transport = default_configure_client_transport;
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
klass->params_set = default_params_set;
klass->params_get = default_params_get;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
klass->make_path_from_uri = default_make_path_from_uri;
rtsp-client: make create_sdp virtual method Fixes https://bugzilla.gnome.org/show_bug.cgi?id=680173
2012-07-18 13:54:49 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
klass->pre_options_request = default_pre_signal_handler;
klass->pre_describe_request = default_pre_signal_handler;
klass->pre_setup_request = default_pre_signal_handler;
klass->pre_play_request = default_pre_signal_handler;
klass->pre_pause_request = default_pre_signal_handler;
klass->pre_teardown_request = default_pre_signal_handler;
klass->pre_set_parameter_request = default_pre_signal_handler;
klass->pre_get_parameter_request = default_pre_signal_handler;
klass->pre_announce_request = default_pre_signal_handler;
klass->pre_record_request = default_pre_signal_handler;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
g_object_class_install_property (gobject_class, PROP_SESSION_POOL,
g_param_spec_object ("session-pool", "Session Pool",
"The session pool to use for client session",
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
GST_TYPE_RTSP_SESSION_POOL,
G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
g_object_class_install_property (gobject_class, PROP_MOUNT_POINTS,
g_param_spec_object ("mount-points", "Mount Points",
"The mount points to use for client session",
GST_TYPE_RTSP_MOUNT_POINTS,
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
g_object_class_install_property (gobject_class, PROP_DROP_BACKLOG,
g_param_spec_boolean ("drop-backlog", "Drop Backlog",
"Drop data when the backlog queue is full",
DEFAULT_DROP_BACKLOG, G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
/**
* GstRTSPClient::post-session-timeout:
*
* An extra tcp timeout ( > 0) after session timeout, in seconds.
* The tcp connection will be kept alive until this timeout happens to give
* the client a possibility to reuse the connection.
* 0 means that the connection will be closed immediately after the session
* timeout.
*
* Default value is -1 seconds, meaning that we let the system close
* the connection.
*
* Since: 1.18
*/
g_object_class_install_property (gobject_class, PROP_POST_SESSION_TIMEOUT,
g_param_spec_int ("post-session-timeout", "Post Session Timeout",
"An extra TCP connection timeout after session timeout", G_MININT,
G_MAXINT, DEFAULT_POST_SESSION_TIMEOUT,
G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
client: emit signal when closing
2011-01-12 14:35:51 +00:00
gst_rtsp_client_signals[SIGNAL_CLOSED] =
g_signal_new ("closed", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
G_STRUCT_OFFSET (GstRTSPClientClass, closed), NULL, NULL, NULL,
G_TYPE_NONE, 0, G_TYPE_NONE);
client: emit signal when closing
2011-01-12 14:35:51 +00:00
add new-session signal to rtsp-client (fixes #683058)
2012-08-30 19:03:27 +00:00
gst_rtsp_client_signals[SIGNAL_NEW_SESSION] =
g_signal_new ("new-session", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
G_STRUCT_OFFSET (GstRTSPClientClass, new_session), NULL, NULL, NULL,
G_TYPE_NONE, 1, GST_TYPE_RTSP_SESSION);
add new-session signal to rtsp-client (fixes #683058)
2012-08-30 19:03:27 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-options-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_OPTIONS_REQUEST] =
g_signal_new ("pre-options-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_options_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::options-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_OPTIONS_REQUEST] =
g_signal_new ("options-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, options_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-describe-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_DESCRIBE_REQUEST] =
g_signal_new ("pre-describe-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_describe_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::describe-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_DESCRIBE_REQUEST] =
g_signal_new ("describe-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, describe_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-setup-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_SETUP_REQUEST] =
g_signal_new ("pre-setup-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_setup_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::setup-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_SETUP_REQUEST] =
g_signal_new ("setup-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, setup_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-play-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_PLAY_REQUEST] =
g_signal_new ("pre-play-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
pre_play_request), pre_signal_accumulator, NULL,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::play-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_PLAY_REQUEST] =
g_signal_new ("play-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, play_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-pause-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_PAUSE_REQUEST] =
g_signal_new ("pre-pause-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_pause_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::pause-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_PAUSE_REQUEST] =
g_signal_new ("pause-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, pause_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-teardown-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_TEARDOWN_REQUEST] =
g_signal_new ("pre-teardown-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_teardown_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::teardown-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_TEARDOWN_REQUEST] =
g_signal_new ("teardown-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, teardown_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-set-parameter-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_SET_PARAMETER_REQUEST] =
g_signal_new ("pre-set-parameter-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_set_parameter_request), pre_signal_accumulator, NULL, NULL,
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::set-parameter-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_SET_PARAMETER_REQUEST] =
g_signal_new ("set-parameter-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
set_parameter_request), NULL, NULL, NULL,
signals: use generic marshal function Use the generic C marshal function. Use more explicit type instead of G_TYPE_POINTER
2014-06-24 07:43:44 +00:00
G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-get-parameter-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_GET_PARAMETER_REQUEST] =
g_signal_new ("pre-get-parameter-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_get_parameter_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::get-parameter-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
gst_rtsp_client_signals[SIGNAL_GET_PARAMETER_REQUEST] =
g_signal_new ("get-parameter-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
get_parameter_request), NULL, NULL, NULL,
signals: use generic marshal function Use the generic C marshal function. Use more explicit type instead of G_TYPE_POINTER
2014-06-24 07:43:44 +00:00
G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::handle-response:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
Add handle-response signal for when we receive a GET_PARAMETER response
2013-08-08 14:57:42 +00:00
gst_rtsp_client_signals[SIGNAL_HANDLE_RESPONSE] =
g_signal_new ("handle-response", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
handle_response), NULL, NULL, NULL,
signals: use generic marshal function Use the generic C marshal function. Use more explicit type instead of G_TYPE_POINTER
2014-06-24 07:43:44 +00:00
G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
Add handle-response signal for when we receive a GET_PARAMETER response
2013-08-08 14:57:42 +00:00
gi: improve annotations Fixes https://bugzilla.gnome.org/show_bug.cgi?id=730953
2014-06-24 07:48:45 +00:00
/**
* GstRTSPClient::send-message:
* @client: The RTSP client
* @session: (type GstRtspServer.RTSPSession): The session
* @message: (type GstRtsp.RTSPMessage): The message
*/
client: emit a signal before sending a message Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728970
2014-05-01 04:17:06 +00:00
gst_rtsp_client_signals[SIGNAL_SEND_MESSAGE] =
g_signal_new ("send-message", G_TYPE_FROM_CLASS (klass),
rtsp-client: Add a send_message default signal handler This allows subclasses to easily hook into the response sending mechanism without doing everything from a signal, which seems awkward from subclasses.
2014-12-29 11:06:50 +00:00
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
send_message), NULL, NULL, NULL,
signals: use generic marshal function Use the generic C marshal function. Use more explicit type instead of G_TYPE_POINTER
2014-06-24 07:43:44 +00:00
G_TYPE_NONE, 2, GST_TYPE_RTSP_CONTEXT, G_TYPE_POINTER);
client: emit a signal before sending a message Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728970
2014-05-01 04:17:06 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-announce-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_ANNOUNCE_REQUEST] =
g_signal_new ("pre-announce-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_announce_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::announce-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
gst_rtsp_client_signals[SIGNAL_ANNOUNCE_REQUEST] =
g_signal_new ("announce-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, announce_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/**
* GstRTSPClient::pre-record-request:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
*
* Returns: a #GstRTSPStatusCode, GST_RTSP_STS_OK in case of success,
* otherwise an appropriate return code
*
* Since: 1.12
*/
gst_rtsp_client_signals[SIGNAL_PRE_RECORD_REQUEST] =
g_signal_new ("pre-record-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
pre_record_request), pre_signal_accumulator, NULL, NULL,
GST_TYPE_RTSP_STATUS_CODE, 1, GST_TYPE_RTSP_CONTEXT);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
/**
* GstRTSPClient::record-request:
* @client: a #GstRTSPClient
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
*/
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
gst_rtsp_client_signals[SIGNAL_RECORD_REQUEST] =
g_signal_new ("record-request", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass, record_request),
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
NULL, NULL, NULL, G_TYPE_NONE, 1, GST_TYPE_RTSP_CONTEXT);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
/**
* GstRTSPClient::check-requirements:
* @client: a #GstRTSPClient
rtsp-client: add type annotations gi doesn't seem to be able to figure out the type of the signal parameters when defined with G_DEFINE_POINTER_TYPE
2018-02-06 16:58:49 +00:00
* @ctx: (type GstRtspServer.RTSPContext): a #GstRTSPContext
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
* @arr: a NULL-terminated array of strings
*
* Returns: a newly allocated string with comma-separated list of
* unsupported options. An empty string must be returned if
* all options are supported.
*
* Since: 1.6
*/
gst_rtsp_client_signals[SIGNAL_CHECK_REQUIREMENTS] =
g_signal_new ("check-requirements", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRTSPClientClass,
Don't pass default GLib marshallers for signals By passing NULL to `g_signal_new` instead of a marshaller, GLib will actually internally optimize the signal (if the marshaller is available in GLib itself) by also setting the valist marshaller. This makes the signal emission a bit more performant than the regular marshalling, which still needs to box into `GValue` and call libffi in case of a generic marshaller. Note that for custom marshallers, one would use `g_signal_set_va_marshaller()` with the valist marshaller instead.
2019-08-29 05:34:26 +00:00
check_requirements), NULL, NULL, NULL,
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
G_TYPE_STRING, 2, GST_TYPE_RTSP_CONTEXT, G_TYPE_STRV);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
tunnels =
g_hash_table_new_full (g_str_hash, g_str_equal, g_free, g_object_unref);
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_init (&tunnels_lock);
Use GStreamer's debugging subsystem
2009-11-21 18:20:23 +00:00
GST_DEBUG_CATEGORY_INIT (rtsp_client_debug, "rtspclient", 0, "GstRTSPClient");
Initial import
2008-10-09 12:29:12 +00:00
}
static void
gst_rtsp_client_init (GstRTSPClient * client)
{
Update for g_type_class_add_private() deprecation in recent GLib
2018-06-24 10:44:26 +00:00
GstRTSPClientPrivate *priv = gst_rtsp_client_get_instance_private (client);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
client->priv = priv;
g_mutex_init (&priv->lock);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_init (&priv->send_lock);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_init (&priv->watch_lock);
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
priv->data_seqs = g_array_new (FALSE, FALSE, sizeof (DataSeq));
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
priv->drop_backlog = DEFAULT_DROP_BACKLOG;
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
priv->post_session_timeout = DEFAULT_POST_SESSION_TIMEOUT;
rtsp: Ref transports in hash table. Also ref streams for transports. This solves a crash when reciving a rtcp after teardown but before client finalize. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=740845
2014-11-28 11:45:14 +00:00
priv->transports =
g_hash_table_new_full (g_direct_hash, g_direct_equal, NULL,
g_object_unref);
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
priv->pipelined_requests = g_hash_table_new_full (g_str_hash,
g_str_equal, g_free, g_free);
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
priv->tstate = TUNNEL_STATE_UNKNOWN;
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
priv->content_length_limit = G_MAXUINT;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
}
static GstRTSPFilterResult
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
filter_session_media (GstRTSPSession * sess, GstRTSPSessionMedia * sessmedia,
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
gpointer user_data)
{
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
gboolean *closed = user_data;
GstRTSPMedia *media;
guint i, n_streams;
gboolean is_all_udp = TRUE;
media = gst_rtsp_session_media_get_media (sessmedia);
n_streams = gst_rtsp_media_n_streams (media);
for (i = 0; i < n_streams; i++) {
GstRTSPStreamTransport *transport =
gst_rtsp_session_media_get_transport (sessmedia, i);
const GstRTSPTransport *rtsp_transport;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
if (!transport)
continue;
rtsp_transport = gst_rtsp_stream_transport_get_transport (transport);
if (rtsp_transport
&& rtsp_transport->lower_transport != GST_RTSP_LOWER_TRANS_UDP
&& rtsp_transport->lower_transport != GST_RTSP_LOWER_TRANS_UDP_MCAST) {
is_all_udp = FALSE;
break;
}
}
if (!is_all_udp || gst_rtsp_media_is_stop_on_disconnect (media)) {
gst_rtsp_session_media_set_state (sessmedia, GST_STATE_NULL);
return GST_RTSP_FILTER_REMOVE;
} else {
*closed = FALSE;
return GST_RTSP_FILTER_KEEP;
}
Initial import
2008-10-09 12:29:12 +00:00
}
client: also watch newly created session When we newly created a session, start watching it immediately instead of on the next request.
2013-06-20 10:20:21 +00:00
static void
client_watch_session (GstRTSPClient * client, GstRTSPSession * session)
{
GstRTSPClientPrivate *priv = client->priv;
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
g_mutex_lock (&priv->lock);
/* check if we already know about this session */
client: fix build
2014-07-01 13:02:15 +00:00
if (g_list_find (priv->sessions, session) == NULL) {
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
GST_INFO ("watching session %p", session);
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
priv->sessions = g_list_prepend (priv->sessions, g_object_ref (session));
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
priv->sessions_cookie++;
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
/* connect removed session handler, it will be disconnected when the last
* session gets removed */
if (priv->session_removed_id == 0)
priv->session_removed_id = g_signal_connect_data (priv->session_pool,
"session-removed", G_CALLBACK (client_session_removed),
g_object_ref (client), (GClosureNotify) g_object_unref, 0);
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
}
g_mutex_unlock (&priv->lock);
client: also watch newly created session When we newly created a session, start watching it immediately instead of on the next request.
2013-06-20 10:20:21 +00:00
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
return;
client: also watch newly created session When we newly created a session, start watching it immediately instead of on the next request.
2013-06-20 10:20:21 +00:00
}
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
/* should be called with lock */
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
static void
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
client_unwatch_session (GstRTSPClient * client, GstRTSPSession * session,
GList * link)
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
{
GstRTSPClientPrivate *priv = client->priv;
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
GST_INFO ("client %p: unwatch session %p", client, session);
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
if (link == NULL) {
link = g_list_find (priv->sessions, session);
if (link == NULL)
return;
}
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
priv->sessions = g_list_delete_link (priv->sessions, link);
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
priv->sessions_cookie++;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
/* if this was the last session, disconnect the handler.
* This will also drop the extra client ref */
if (!priv->sessions) {
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
g_signal_handler_disconnect (priv->session_pool, priv->session_removed_id);
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
priv->session_removed_id = 0;
}
rtsp-client: Session filter in unwatch session Call session filter with filter_session_media as paramer in client_unwatch_session if using drop_backlog = FALSE. In client_unwatch_session its allowed to grow the watchs backlog. If using drop_backlog = FALSE and the backlog is full it will cause a deadlock when setting session media state to NULL if the backlog is not allowed to grow. https://bugzilla.gnome.org/show_bug.cgi?id=771983
2016-09-26 09:16:04 +00:00
if (!priv->drop_backlog) {
/* unlink all media managed in this session */
gst_rtsp_session_filter (session, filter_session_media, client);
}
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
/* remove the session */
g_object_unref (session);
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
}
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
static GstRTSPFilterResult
cleanup_session (GstRTSPClient * client, GstRTSPSession * sess,
gpointer user_data)
client: improve client cleanups Make sure the session does not timeout when using TCP. We need to do this because quicktime player does not send RTCP for some reason in tunneled mode. Refactor some cleanup code. Fixes #612915
2010-04-06 15:08:40 +00:00
{
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
gboolean *closed = user_data;
rtsp-client: Session filter in unwatch session Call session filter with filter_session_media as paramer in client_unwatch_session if using drop_backlog = FALSE. In client_unwatch_session its allowed to grow the watchs backlog. If using drop_backlog = FALSE and the backlog is full it will cause a deadlock when setting session media state to NULL if the backlog is not allowed to grow. https://bugzilla.gnome.org/show_bug.cgi?id=771983
2016-09-26 09:16:04 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
rtsp-client: Session filter in unwatch session Call session filter with filter_session_media as paramer in client_unwatch_session if using drop_backlog = FALSE. In client_unwatch_session its allowed to grow the watchs backlog. If using drop_backlog = FALSE and the backlog is full it will cause a deadlock when setting session media state to NULL if the backlog is not allowed to grow. https://bugzilla.gnome.org/show_bug.cgi?id=771983
2016-09-26 09:16:04 +00:00
if (priv->drop_backlog) {
/* unlink all media managed in this session. This needs to happen
* without the client lock, so we really want to do it here. */
gst_rtsp_session_filter (sess, filter_session_media, user_data);
}
client: set session media to NULL without the lock We need to set session medias to NULL without the client lock otherwise we can end up in a deadlock if another thread is waiting for the lock and media unprepare is also waiting for that thread to end. https://bugzilla.gnome.org/show_bug.cgi?id=737690
2014-09-30 23:36:51 +00:00
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
if (*closed)
return GST_RTSP_FILTER_REMOVE;
else
return GST_RTSP_FILTER_KEEP;
client: improve client cleanups Make sure the session does not timeout when using TCP. We need to do this because quicktime player does not send RTCP for some reason in tunneled mode. Refactor some cleanup code. Fixes #612915
2010-04-06 15:08:40 +00:00
}
client: refactor cleanup of cached media
2014-11-07 11:48:53 +00:00
static void
clean_cached_media (GstRTSPClient * client, gboolean unprepare)
{
GstRTSPClientPrivate *priv = client->priv;
if (priv->path) {
g_free (priv->path);
priv->path = NULL;
}
if (priv->media) {
if (unprepare)
gst_rtsp_media_unprepare (priv->media);
g_object_unref (priv->media);
priv->media = NULL;
}
}
client: improve client cleanups Make sure the session does not timeout when using TCP. We need to do this because quicktime player does not send RTCP for some reason in tunneled mode. Refactor some cleanup code. Fixes #612915
2010-04-06 15:08:40 +00:00
/* A client is finalized when the connection is broken */
static void
gst_rtsp_client_finalize (GObject * obj)
{
GstRTSPClient *client = GST_RTSP_CLIENT (obj);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
client: improve client cleanups Make sure the session does not timeout when using TCP. We need to do this because quicktime player does not send RTCP for some reason in tunneled mode. Refactor some cleanup code. Fixes #612915
2010-04-06 15:08:40 +00:00
GST_INFO ("finalize client %p", client);
rtsp-client: Clean up watch/watch context and related state consistently And assert that it was cleaned up properly before the client is finalized. If something is still around when the client is shut down then something went very wrong before. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:16:18 +00:00
/* the watch and related state should be cleared before finalize
* as the watch actually holds a strong reference to the client */
g_assert (priv->watch == NULL);
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
g_assert (priv->rtsp_ctrl_timeout == NULL);
rtsp-client: Use watch_context before unref Move the usage of priv->watch_context to beginning of function gst_rtsp_client_finalize. Instead of use it after g_main_context_unref (priv->watch_context).
2020-03-09 13:17:34 +00:00
rtsp-client: Only unref client watch context on finalize, to avoid deadlock Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/issues/127 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/176>
2020-12-14 01:12:38 +00:00
if (priv->watch_context) {
g_main_context_unref (priv->watch_context);
priv->watch_context = NULL;
}
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_client_set_send_messages_func (client, NULL, NULL, NULL);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
/* all sessions should have been removed by now. We keep a ref to
* the client object for the session removed handler. The ref is
* dropped when the last session is removed from the list. */
g_assert (priv->sessions == NULL);
g_assert (priv->session_removed_id == 0);
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
g_array_unref (priv->data_seqs);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
g_hash_table_unref (priv->transports);
rtsp-client: unref 'pipelined_requests' in finalize The hash table priv->pipelined_requests is not unref:ed in the finalize funktion. Make sure it is. https://bugzilla.gnome.org/show_bug.cgi?id=788704
2017-10-09 10:43:01 +00:00
g_hash_table_unref (priv->pipelined_requests);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (priv->connection)
gst_rtsp_connection_free (priv->connection);
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
if (priv->session_pool) {
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_object_unref (priv->session_pool);
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
}
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (priv->mount_points)
g_object_unref (priv->mount_points);
if (priv->auth)
g_object_unref (priv->auth);
client: free threadpool https://bugzilla.gnome.org/show_bug.cgi?id=707638
2013-09-05 06:56:02 +00:00
if (priv->thread_pool)
g_object_unref (priv->thread_pool);
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
client: refactor cleanup of cached media
2014-11-07 11:48:53 +00:00
clean_cached_media (client, TRUE);
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_free (priv->server_ip);
g_mutex_clear (&priv->lock);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_clear (&priv->send_lock);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_clear (&priv->watch_lock);
client: use right size for malloc
2010-03-16 17:33:23 +00:00
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
G_OBJECT_CLASS (gst_rtsp_client_parent_class)->finalize (obj);
}
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
static void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
gst_rtsp_client_get_property (GObject * object, guint propid,
GValue * value, GParamSpec * pspec)
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
{
GstRTSPClient *client = GST_RTSP_CLIENT (object);
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
GstRTSPClientPrivate *priv = client->priv;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
switch (propid) {
case PROP_SESSION_POOL:
g_value_take_object (value, gst_rtsp_client_get_session_pool (client));
break;
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
case PROP_MOUNT_POINTS:
g_value_take_object (value, gst_rtsp_client_get_mount_points (client));
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
break;
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
case PROP_DROP_BACKLOG:
g_value_set_boolean (value, priv->drop_backlog);
break;
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
case PROP_POST_SESSION_TIMEOUT:
g_value_set_int (value, priv->post_session_timeout);
break;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, propid, pspec);
}
}
static void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
gst_rtsp_client_set_property (GObject * object, guint propid,
const GValue * value, GParamSpec * pspec)
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
{
GstRTSPClient *client = GST_RTSP_CLIENT (object);
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
GstRTSPClientPrivate *priv = client->priv;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
switch (propid) {
case PROP_SESSION_POOL:
gst_rtsp_client_set_session_pool (client, g_value_get_object (value));
break;
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
case PROP_MOUNT_POINTS:
gst_rtsp_client_set_mount_points (client, g_value_get_object (value));
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
break;
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
case PROP_DROP_BACKLOG:
g_mutex_lock (&priv->lock);
priv->drop_backlog = g_value_get_boolean (value);
g_mutex_unlock (&priv->lock);
break;
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
case PROP_POST_SESSION_TIMEOUT:
g_mutex_lock (&priv->lock);
priv->post_session_timeout = g_value_get_int (value);
g_mutex_unlock (&priv->lock);
break;
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, propid, pspec);
}
}
Initial import
2008-10-09 12:29:12 +00:00
/**
* gst_rtsp_client_new:
*
* Create a new #GstRTSPClient instance.
docs: improve docs
2010-12-30 11:41:31 +00:00
*
gobject-introspection: Add annotations to support language bindings In addition a few cosmetic changes: * Adjust the order of arguments * Fix typo: occured -> occurred * Fix indentation after Return:-clauses Fixes https://bugzilla.gnome.org/show_bug.cgi?id=726941
2014-03-23 20:24:48 +00:00
* Returns: (transfer full): a new #GstRTSPClient
Initial import
2008-10-09 12:29:12 +00:00
*/
GstRTSPClient *
gst_rtsp_client_new (void)
{
GstRTSPClient *result;
result = g_object_new (GST_TYPE_RTSP_CLIENT, NULL);
return result;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
static void
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (GstRTSPClient * client, GstRTSPContext * ctx,
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
GstRTSPMessage * message, gboolean close)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_add_header (message, GST_RTSP_HDR_SERVER,
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
"GStreamer RTSP server");
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
/* remove any previous header */
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_remove_header (message, GST_RTSP_HDR_SESSION, -1);
Add timeout property Add a timeout property ot the client and make the other properties into GObject properties.
2009-02-10 15:24:13 +00:00
Some more session timeout handling Move the session header setting code to a central place so that we always add the timeout parameter too. Handle timeouts by running the session cleanup code. Stop media before cleaning up.
2009-02-13 11:57:45 +00:00
/* add the new session header for new session ids */
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
if (ctx->session) {
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_take_header (message, GST_RTSP_HDR_SESSION,
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
gst_rtsp_session_get_header (ctx->session));
Some more session timeout handling Move the session header setting code to a central place so that we always add the timeout parameter too. Handle timeouts by running the session cleanup code. Stop media before cleaning up.
2009-02-13 11:57:45 +00:00
}
client: dump rtsp message only if debug threshold is higher than GST_LEVEL_LOG
2009-11-21 00:17:25 +00:00
if (gst_debug_category_get_threshold (rtsp_client_debug) >= GST_LEVEL_LOG) {
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_dump (message);
client: dump rtsp message only if debug threshold is higher than GST_LEVEL_LOG
2009-11-21 00:17:25 +00:00
}
Some more session timeout handling Move the session header setting code to a central place so that we always add the timeout parameter too. Handle timeouts by running the session cleanup code. Stop media before cleaning up.
2009-02-13 11:57:45 +00:00
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
if (close)
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_add_header (message, GST_RTSP_HDR_CONNECTION, "close");
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (ctx->request)
message->type_data.response.version =
ctx->request->type_data.request.version;
client: emit a signal before sending a message Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728970
2014-05-01 04:17:06 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_SEND_MESSAGE],
0, ctx, message);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_lock (&priv->send_lock);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
if (priv->send_messages_func) {
priv->send_messages_func (client, message, 1, close, priv->send_data);
} else if (priv->send_func) {
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
priv->send_func (client, message, close, priv->send_data);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
}
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_unlock (&priv->send_lock);
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
gst_rtsp_message_unset (message);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
}
Initial import
2008-10-09 12:29:12 +00:00
static void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
send_generic_response (GstRTSPClient * client, GstRTSPStatusCode code,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext * ctx)
Initial import
2008-10-09 12:29:12 +00:00
{
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
ctx->session = NULL;
send_message (client, ctx, ctx->response, FALSE);
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
}
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
static void
send_option_not_supported_response (GstRTSPClient * client,
GstRTSPContext * ctx, const gchar * unsupported_options)
{
GstRTSPStatusCode code = GST_RTSP_STS_OPTION_NOT_SUPPORTED;
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
if (unsupported_options != NULL) {
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_UNSUPPORTED,
unsupported_options);
}
ctx->session = NULL;
send_message (client, ctx, ctx->response, FALSE);
}
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
static gboolean
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
paths_are_equal (const gchar * path1, const gchar * path2, gint len2)
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
{
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (path1 == NULL || path2 == NULL)
return FALSE;
if (strlen (path1) != len2)
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
return FALSE;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (strncmp (path1, path2, len2))
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
return FALSE;
return TRUE;
Initial import
2008-10-09 12:29:12 +00:00
}
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
/* this function is called to initially find the media for the DESCRIBE request
* but is cached for when the same client (without breaking the connection) is
* doing a setup for the exact same url. */
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
static GstRTSPMedia *
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
find_media (GstRTSPClient * client, GstRTSPContext * ctx, gchar * path,
gint * matched)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
GstRTSPMediaFactory *factory;
GstRTSPMedia *media;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
gint path_len;
/* find the longest matching factory for the uri first */
if (!(factory = gst_rtsp_mount_points_match (priv->mount_points,
path, matched)))
goto no_factory;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->factory = factory;
auth: add auth checks Add an enum with auth checks and implement the checks in the auth object. Perform the checks from the client.
2013-07-08 14:29:01 +00:00
auth: simplify auth checks Remove client from methods, it's now in the state Perform the check specified by the string, use the information from the thread local context.
2013-07-09 14:04:35 +00:00
if (!gst_rtsp_auth_check (GST_RTSP_AUTH_CHECK_MEDIA_FACTORY_ACCESS))
goto no_factory_access;
if (!gst_rtsp_auth_check (GST_RTSP_AUTH_CHECK_MEDIA_FACTORY_CONSTRUCT))
auth: add auth checks Add an enum with auth checks and implement the checks in the auth object. Perform the checks from the client.
2013-07-08 14:29:01 +00:00
goto not_authorized;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (matched)
path_len = *matched;
else
path_len = strlen (path);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (!paths_are_equal (priv->path, path, path_len)) {
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
/* remove any previously cached values before we try to construct a new
* media for uri */
client: refactor cleanup of cached media
2014-11-07 11:48:53 +00:00
clean_cached_media (client, TRUE);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
/* prepare the media and add it to the pipeline */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!(media = gst_rtsp_media_factory_construct (factory, ctx->uri)))
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
goto no_media;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->media = media;
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if (!(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_RECORD)) {
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
GstRTSPThread *thread;
thread = gst_rtsp_thread_pool_get_thread (priv->thread_pool,
GST_RTSP_THREAD_TYPE_MEDIA, ctx);
if (thread == NULL)
goto no_thread;
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
/* prepare the media */
if (!gst_rtsp_media_prepare (media, thread))
goto no_prepare;
}
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
/* now keep track of the uri and the media */
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
priv->path = g_strndup (path, path_len);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->media = media;
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
} else {
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
/* we have seen this path before, used cached media */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
media = priv->media;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->media = media;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
GST_INFO ("reusing cached media %p for path %s", media, priv->path);
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
g_object_unref (factory);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->factory = NULL;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
if (media)
g_object_ref (media);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
return media;
/* ERRORS */
no_factory:
{
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
GST_ERROR ("client %p: no factory for path %s", client, path);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
return NULL;
}
auth: simplify auth checks Remove client from methods, it's now in the state Perform the check specified by the string, use the information from the thread local context.
2013-07-09 14:04:35 +00:00
no_factory_access:
{
rtsp-client: Fix factory leaking in find_media() in error cases https://bugzilla.gnome.org/show_bug.cgi?id=771488
2016-09-14 14:48:39 +00:00
g_object_unref (factory);
ctx->factory = NULL;
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
GST_ERROR ("client %p: not authorized to see factory path %s", client,
path);
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
/* error reply is already sent */
auth: simplify auth checks Remove client from methods, it's now in the state Perform the check specified by the string, use the information from the thread local context.
2013-07-09 14:04:35 +00:00
return NULL;
}
auth: add auth checks Add an enum with auth checks and implement the checks in the auth object. Perform the checks from the client.
2013-07-08 14:29:01 +00:00
not_authorized:
{
rtsp-client: Fix factory leaking in find_media() in error cases https://bugzilla.gnome.org/show_bug.cgi?id=771488
2016-09-14 14:48:39 +00:00
g_object_unref (factory);
ctx->factory = NULL;
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
GST_ERROR ("client %p: not authorized for factory path %s", client, path);
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
/* error reply is already sent */
auth: add auth checks Add an enum with auth checks and implement the checks in the auth object. Perform the checks from the client.
2013-07-08 14:29:01 +00:00
return NULL;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_media:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: can't create media", client);
client: Reply 400 if media cannot be constructed Reply 400 Bad Request instead of 503 Service Unavailable if media cannot be constructed in SETUP. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708821
2013-09-26 09:20:05 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
g_object_unref (factory);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->factory = NULL;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
return NULL;
}
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
no_thread:
{
GST_ERROR ("client %p: can't create thread", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
g_object_unref (media);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->media = NULL;
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
g_object_unref (factory);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->factory = NULL;
client: pass thread from pool to media _prepare Get a thread from the configured threadpool and pass it to the prepare method of the media.
2013-07-10 15:09:27 +00:00
return NULL;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_prepare:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: can't prepare media", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
g_object_unref (media);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->media = NULL;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
g_object_unref (factory);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->factory = NULL;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
return NULL;
}
}
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
static inline DataSeq *
get_data_seq_element (GstRTSPClient * client, guint8 channel)
{
GstRTSPClientPrivate *priv = client->priv;
GArray *data_seqs = priv->data_seqs;
gint i = 0;
while (i < data_seqs->len) {
DataSeq *data_seq = &g_array_index (data_seqs, DataSeq, i);
if (data_seq->channel == channel)
return data_seq;
i++;
}
return NULL;
}
static void
add_data_seq (GstRTSPClient * client, guint8 channel)
{
GstRTSPClientPrivate *priv = client->priv;
DataSeq data_seq = {.channel = channel,.seq = 0 };
if (get_data_seq_element (client, channel) == NULL)
g_array_append_val (priv->data_seqs, data_seq);
}
static void
set_data_seq (GstRTSPClient * client, guint8 channel, guint seq)
{
DataSeq *data_seq;
data_seq = get_data_seq_element (client, channel);
g_assert_nonnull (data_seq);
data_seq->seq = seq;
}
static guint
get_data_seq (GstRTSPClient * client, guint8 channel)
{
DataSeq *data_seq;
data_seq = get_data_seq_element (client, channel);
g_assert_nonnull (data_seq);
return data_seq->seq;
}
static gboolean
get_data_channel (GstRTSPClient * client, guint seq, guint8 * channel)
{
GstRTSPClientPrivate *priv = client->priv;
GArray *data_seqs = priv->data_seqs;
gint i = 0;
while (i < data_seqs->len) {
DataSeq *data_seq = &g_array_index (data_seqs, DataSeq, i);
if (data_seq->seq == seq) {
*channel = data_seq->channel;
return TRUE;
}
i++;
}
return FALSE;
}
static gboolean
do_close (gpointer user_data)
{
GstRTSPClient *client = user_data;
gst_rtsp_client_close (client);
return G_SOURCE_REMOVE;
}
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
static gboolean
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
do_send_data (GstBuffer * buffer, guint8 channel, GstRTSPClient * client)
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
GstRTSPMessage message = { 0 };
rtsp-client: Fix session timeout When streaming data over TCP then is not the keep-alive functionality working. The reason is that the function do_send_data have changed to boolean but the code is still checking the received result from send_func with GST_RTSP_OK. The result is that a successful send_func will always lead to that do_send_data is returning false and the keep-alive will not be updated. https://bugzilla.gnome.org/show_bug.cgi?id=795321
2018-04-17 09:03:11 +00:00
gboolean ret = TRUE;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
gst_rtsp_message_init_data (&message, channel);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_message_set_body_buffer (&message, buffer);
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_lock (&priv->send_lock);
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
if (get_data_seq (client, channel) != 0) {
GST_WARNING ("already a queued data message for channel %d", channel);
g_mutex_unlock (&priv->send_lock);
return FALSE;
}
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
if (priv->send_messages_func) {
ret =
priv->send_messages_func (client, &message, 1, FALSE, priv->send_data);
} else if (priv->send_func) {
rtsp-client: Fix session timeout When streaming data over TCP then is not the keep-alive functionality working. The reason is that the function do_send_data have changed to boolean but the code is still checking the received result from send_func with GST_RTSP_OK. The result is that a successful send_func will always lead to that do_send_data is returning false and the keep-alive will not be updated. https://bugzilla.gnome.org/show_bug.cgi?id=795321
2018-04-17 09:03:11 +00:00
ret = priv->send_func (client, &message, FALSE, priv->send_data);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
}
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_unlock (&priv->send_lock);
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
gst_rtsp_message_unset (&message);
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
if (!ret) {
GSource *idle_src;
/* close in watch context */
idle_src = g_idle_source_new ();
g_source_set_callback (idle_src, do_close, client, NULL);
g_source_attach (idle_src, priv->watch_context);
g_source_unref (idle_src);
}
rtsp-client: Fix session timeout When streaming data over TCP then is not the keep-alive functionality working. The reason is that the function do_send_data have changed to boolean but the code is still checking the received result from send_func with GST_RTSP_OK. The result is that a successful send_func will always lead to that do_send_data is returning false and the keep-alive will not be updated. https://bugzilla.gnome.org/show_bug.cgi?id=795321
2018-04-17 09:03:11 +00:00
return ret;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
}
stream: refactor TCP backpressure handling The previous implementation stopped sending TCP messages to all clients when a single one stopped consuming them, which obviously created problems for shared media. Instead, we now manage a backlog in stream-transport, and slow clients are removed once this backlog exceeds a maximum duration, currently hardcoded. Fixes #80
2019-10-15 17:08:32 +00:00
static gboolean
do_check_back_pressure (guint8 channel, GstRTSPClient * client)
{
return get_data_seq (client, channel) != 0;
}
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
static gboolean
do_send_data_list (GstBufferList * buffer_list, guint8 channel,
GstRTSPClient * client)
{
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
gboolean ret = TRUE;
guint i, n = gst_buffer_list_length (buffer_list);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
GstRTSPMessage *messages;
g_mutex_lock (&priv->send_lock);
if (get_data_seq (client, channel) != 0) {
GST_WARNING ("already a queued data message for channel %d", channel);
g_mutex_unlock (&priv->send_lock);
return FALSE;
}
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
messages = g_newa (GstRTSPMessage, n);
memset (messages, 0, sizeof (GstRTSPMessage) * n);
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
for (i = 0; i < n; i++) {
GstBuffer *buffer = gst_buffer_list_get (buffer_list, i);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_message_init_data (&messages[i], channel);
gst_rtsp_message_set_body_buffer (&messages[i], buffer);
}
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
if (priv->send_messages_func) {
ret =
priv->send_messages_func (client, messages, n, FALSE, priv->send_data);
} else if (priv->send_func) {
for (i = 0; i < n; i++) {
ret = priv->send_func (client, &messages[i], FALSE, priv->send_data);
if (!ret)
break;
}
}
g_mutex_unlock (&priv->send_lock);
for (i = 0; i < n; i++) {
gst_rtsp_message_unset (&messages[i]);
}
if (!ret) {
GSource *idle_src;
/* close in watch context */
idle_src = g_idle_source_new ();
g_source_set_callback (idle_src, do_close, client, NULL);
g_source_attach (idle_src, priv->watch_context);
g_source_unref (idle_src);
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
}
return ret;
}
client: expose _close() method Expose a previously internal close method to close the client connection.
2014-07-10 15:05:13 +00:00
/**
* gst_rtsp_client_close:
* @client: a #GstRTSPClient
*
* Close the connection of @client and remove all media it was managing.
*
* Since: 1.4
*/
void
gst_rtsp_client_close (GstRTSPClient * client)
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-server: Run gst-indent Since it wasn't using the upstream common previously, there was no indentation check before commiting.
2010-12-11 09:48:42 +00:00
const gchar *tunnelid;
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
GST_DEBUG ("client %p: closing connection", client);
rtsp-client: make closing more thread safe + Take the watch lock prior to using priv->watch + Flush both the watch and connection before closing / unreffing gst_rtsp_connection_close() is not threadsafe on its own, this is a workaround at the client level, where we control both the watch and the connection
2020-01-15 16:06:41 +00:00
g_mutex_lock (&priv->watch_lock);
/* Work around the lack of thread safety of gst_rtsp_connection_close */
if (priv->watch) {
gst_rtsp_watch_set_flushing (priv->watch, TRUE);
}
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
if (priv->connection) {
if ((tunnelid = gst_rtsp_connection_get_tunnelid (priv->connection))) {
g_mutex_lock (&tunnels_lock);
/* remove from tunnelids */
g_hash_table_remove (tunnels, tunnelid);
g_mutex_unlock (&tunnels_lock);
}
rtsp-client: make closing more thread safe + Take the watch lock prior to using priv->watch + Flush both the watch and connection before closing / unreffing gst_rtsp_connection_close() is not threadsafe on its own, this is a workaround at the client level, where we control both the watch and the connection
2020-01-15 16:06:41 +00:00
gst_rtsp_connection_flush (priv->connection, TRUE);
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
gst_rtsp_connection_close (priv->connection);
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
}
client: Destroy the rtsp watch after connection close
2014-06-12 11:48:44 +00:00
if (priv->watch) {
GST_DEBUG ("client %p: destroying watch", client);
g_source_destroy ((GSource *) priv->watch);
priv->watch = NULL;
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_client_set_send_messages_func (client, NULL, NULL, NULL);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove (client);
rtsp-client: Clean up watch/watch context and related state consistently And assert that it was cleaned up properly before the client is finalized. If something is still around when the client is shut down then something went very wrong before. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:16:18 +00:00
}
rtsp-client: make closing more thread safe + Take the watch lock prior to using priv->watch + Flush both the watch and connection before closing / unreffing gst_rtsp_connection_close() is not threadsafe on its own, this is a workaround at the client level, where we control both the watch and the connection
2020-01-15 16:06:41 +00:00
g_mutex_unlock (&priv->watch_lock);
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
}
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
static gchar *
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
default_make_path_from_uri (GstRTSPClient * client, const GstRTSPUrl * uri)
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
{
gchar *path;
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
if (uri->query) {
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
path = g_strconcat (uri->abspath, "?", uri->query, NULL);
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
} else {
/* normalize rtsp://<IP>:<PORT> to rtsp://<IP>:<PORT>/ */
path = g_strdup (uri->abspath[0] ? uri->abspath : "/");
}
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
return path;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/* Default signal handler function for all "pre-command" signals, like
* pre-options-request. It just returns the RTSP return code 200.
* Subclasses can override this to get another default behaviour.
*/
static GstRTSPStatusCode
default_pre_signal_handler (GstRTSPClient * client, GstRTSPContext * ctx)
{
GST_LOG_OBJECT (client, "returning GST_RTSP_STS_OK");
return GST_RTSP_STS_OK;
}
/* The pre-signal accumulator function checks the return value of the signal
* handlers. If any of them returns an RTSP status code that does not start
* with 2 it will return FALSE, no more signal handlers will be called, and
* this last RTSP status code will be the result of the signal emission.
*/
static gboolean
pre_signal_accumulator (GSignalInvocationHint * ihint, GValue * return_accu,
const GValue * handler_return, gpointer data)
{
GstRTSPStatusCode handler_value = g_value_get_enum (handler_return);
GstRTSPStatusCode accumulated_value = g_value_get_enum (return_accu);
if (handler_value < 200 || handler_value > 299) {
GST_DEBUG ("handler_value : %d, returning FALSE", handler_value);
g_value_set_enum (return_accu, handler_value);
return FALSE;
}
/* the accumulated value is initiated to 0 by GLib. if current handler value is
* bigger then use that instead
*
* FIXME: Should we prioritize the 2xx codes in a smarter way?
* Like, "201 Created" > "250 Low On Storage Space" > "200 OK"?
*/
if (handler_value > accumulated_value)
g_value_set_enum (return_accu, handler_value);
return TRUE;
}
rtsp-client: cleanup transports during TEARDOWN When tunneling RTP over RTSP the stream transports are stored in a hash table in the GstRTSPClientPrivate struct. They are used for, among other things, mapping channel id to stream transports when receiving data from the client. The stream tranports are created and added to the hash table in handle_setup_request(), but unfortuately they are not removed in handle_teardown_request(). This means that if the client sends data on the RTSP connection after it has sent the TEARDOWN, which is often the case when audio backchannel is enabled, handle_data() will still be able to map the channel to a session transport and pass the data along to it. Which eventually leads to a failing assert in gst_rtsp_stream_recv_rtp() because the stream is no longer joined to a bin. We avoid this by removing the stream transports from the hash table when we handle the TEARDOWN request. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/184>
2021-01-22 07:58:23 +00:00
/* The cleanup_transports function is called from handle_teardown_request() to
* remove any stream transports from the newly closed session that were added to
* priv->transports in handle_setup_request(). This is done to avoid forwarding
* data from the client on a channel that we just closed.
*/
static void
cleanup_transports (GstRTSPClient * client, GPtrArray * transports)
{
GstRTSPClientPrivate *priv = client->priv;
GstRTSPStreamTransport *stream_transport;
const GstRTSPTransport *rtsp_transport;
guint i;
GST_LOG_OBJECT (client, "potentially removing %u transports",
transports->len);
for (i = 0; i < transports->len; i++) {
stream_transport = g_ptr_array_index (transports, i);
if (stream_transport == NULL) {
GST_LOG_OBJECT (client, "stream transport %u is NULL, continue", i);
continue;
}
rtsp_transport = gst_rtsp_stream_transport_get_transport (stream_transport);
if (rtsp_transport == NULL) {
GST_LOG_OBJECT (client, "RTSP transport %u is NULL, continue", i);
continue;
}
/* priv->transport only stores transports where RTP is tunneled over RTSP */
if (rtsp_transport->lower_transport == GST_RTSP_LOWER_TRANS_TCP) {
if (!g_hash_table_remove (priv->transports,
GINT_TO_POINTER (rtsp_transport->interleaved.min))) {
GST_WARNING_OBJECT (client,
"failed removing transport with key '%d' from priv->transports",
rtsp_transport->interleaved.min);
}
if (!g_hash_table_remove (priv->transports,
GINT_TO_POINTER (rtsp_transport->interleaved.max))) {
GST_WARNING_OBJECT (client,
"failed removing transport with key '%d' from priv->transports",
rtsp_transport->interleaved.max);
}
} else {
GST_LOG_OBJECT (client, "transport %u not RTP/RTSP, skip it", i);
}
}
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_teardown_request (GstRTSPClient * client, GstRTSPContext * ctx)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
GstRTSPClientClass *klass;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
GstRTSPSession *session;
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
GstRTSPSessionMedia *sessmedia;
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
GstRTSPMedia *media;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
GstRTSPStatusCode code;
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
gchar *path;
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
gint matched;
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
gboolean keep_session;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
rtsp-client: cleanup transports during TEARDOWN When tunneling RTP over RTSP the stream transports are stored in a hash table in the GstRTSPClientPrivate struct. They are used for, among other things, mapping channel id to stream transports when receiving data from the client. The stream tranports are created and added to the hash table in handle_setup_request(), but unfortuately they are not removed in handle_teardown_request(). This means that if the client sends data on the RTSP connection after it has sent the TEARDOWN, which is often the case when audio backchannel is enabled, handle_data() will still be able to map the channel to a session transport and pass the data along to it. Which eventually leads to a failing assert in gst_rtsp_stream_recv_rtp() because the stream is no longer joined to a bin. We avoid this by removing the stream transports from the hash table when we handle the TEARDOWN request. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/184>
2021-01-22 07:58:23 +00:00
GPtrArray *session_media_transports;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
Revert "client: map URL to path in requests" This reverts commit e3fded2cec897a2ec003450607b916cc1601fd2d. This is not correct, we only remap the URL to a path in DESCRIBE, the SDP then contains the base and control urls which are used in the SETUP, PLAY, PAUSE and TEARDOWN requests.
2013-09-17 09:41:57 +00:00
if (!ctx->session)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
goto no_session;
Revert "client: map URL to path in requests" This reverts commit e3fded2cec897a2ec003450607b916cc1601fd2d. This is not correct, we only remap the URL to a path in DESCRIBE, the SDP then contains the base and control urls which are used in the SETUP, PLAY, PAUSE and TEARDOWN requests.
2013-09-17 09:41:57 +00:00
session = ctx->session;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!ctx->uri)
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
goto no_uri;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
path = klass->make_path_from_uri (client, ctx->uri);
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* get a handle to the configuration of the media in the session */
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
sessmedia = gst_rtsp_session_dup_media (session, path, &matched);
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
if (!sessmedia)
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
goto not_found;
Initial import
2008-10-09 12:29:12 +00:00
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
/* only aggregate control for now.. */
if (path[matched] != '\0')
goto no_aggregate;
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->sessmedia = sessmedia;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
media = gst_rtsp_session_media_get_media (sessmedia);
g_object_ref (media);
gst_rtsp_media_lock (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_TEARDOWN_REQUEST],
0, ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp-client: cleanup transports during TEARDOWN When tunneling RTP over RTSP the stream transports are stored in a hash table in the GstRTSPClientPrivate struct. They are used for, among other things, mapping channel id to stream transports when receiving data from the client. The stream tranports are created and added to the hash table in handle_setup_request(), but unfortuately they are not removed in handle_teardown_request(). This means that if the client sends data on the RTSP connection after it has sent the TEARDOWN, which is often the case when audio backchannel is enabled, handle_data() will still be able to map the channel to a session transport and pass the data along to it. Which eventually leads to a failing assert in gst_rtsp_stream_recv_rtp() because the stream is no longer joined to a bin. We avoid this by removing the stream transports from the hash table when we handle the TEARDOWN request. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/184>
2021-01-22 07:58:23 +00:00
/* get a reference to the transports in the session media so we can clean up
* our priv->transports before returning */
session_media_transports = gst_rtsp_session_media_get_transports (sessmedia);
client: send out teardown signal before tearing down The advantage is that in the signal handler you get direct access to information about what streams are about to get torn down (in the GstRTSPClientState). Fixes https://bugzilla.gnome.org/show_bug.cgi?id=697686
2013-04-15 10:21:54 +00:00
/* we emit the signal before closing the connection */
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_TEARDOWN_REQUEST],
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
0, ctx);
client: send out teardown signal before tearing down The advantage is that in the signal handler you get direct access to information about what streams are about to get torn down (in the GstRTSPClientState). Fixes https://bugzilla.gnome.org/show_bug.cgi?id=697686
2013-04-15 10:21:54 +00:00
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
gst_rtsp_session_media_set_state (sessmedia, GST_STATE_NULL);
Initial import
2008-10-09 12:29:12 +00:00
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
/* unmanage the media in the session, returns false if all media session
* are torn down. */
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
keep_session = gst_rtsp_session_release_media (session, sessmedia);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* construct the response now */
code = GST_RTSP_STS_OK;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, TRUE);
Some more session timeout handling Move the session header setting code to a central place so that we always add the timeout parameter too. Handle timeouts by running the session cleanup code. Stop media before cleaning up.
2009-02-13 11:57:45 +00:00
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
if (!keep_session) {
/* remove the session */
gst_rtsp_session_pool_remove (priv->session_pool, session);
}
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: cleanup transports during TEARDOWN When tunneling RTP over RTSP the stream transports are stored in a hash table in the GstRTSPClientPrivate struct. They are used for, among other things, mapping channel id to stream transports when receiving data from the client. The stream tranports are created and added to the hash table in handle_setup_request(), but unfortuately they are not removed in handle_teardown_request(). This means that if the client sends data on the RTSP connection after it has sent the TEARDOWN, which is often the case when audio backchannel is enabled, handle_data() will still be able to map the channel to a session transport and pass the data along to it. Which eventually leads to a failing assert in gst_rtsp_stream_recv_rtp() because the stream is no longer joined to a bin. We avoid this by removing the stream transports from the hash table when we handle the TEARDOWN request. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/184>
2021-01-22 07:58:23 +00:00
/* remove all transports that were present in the session media which we just
* unmanaged from the priv->transports array, so we do not try to handle data
* on channels that were just closed */
cleanup_transports (client, session_media_transports);
g_ptr_array_unref (session_media_transports);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
Initial import
2008-10-09 12:29:12 +00:00
/* ERRORS */
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_session:
Initial import
2008-10-09 12:29:12 +00:00
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no session", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
Initial import
2008-10-09 12:29:12 +00:00
return FALSE;
}
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
no_uri:
{
GST_ERROR ("client %p: no uri supplied", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
return FALSE;
}
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
not_found:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no media for uri", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
return FALSE;
}
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
no_aggregate:
{
GST_ERROR ("client %p: no aggregate path %s", client, path);
send_generic_response (client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GST_RTSP_STS_ONLY_AGGREGATE_OPERATION_ALLOWED, ctx);
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
return FALSE;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
return FALSE;
}
Initial import
2008-10-09 12:29:12 +00:00
}
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
static GstRTSPResult
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
default_params_set (GstRTSPClient * client, GstRTSPContext * ctx)
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
{
GstRTSPResult res;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = gst_rtsp_params_set (client, ctx);
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
return res;
}
static GstRTSPResult
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
default_params_get (GstRTSPClient * client, GstRTSPContext * ctx)
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
{
GstRTSPResult res;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = gst_rtsp_params_get (client, ctx);
rtsp-client: Make param_set and param_get virtual Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702072
2013-06-12 10:23:56 +00:00
return res;
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_get_param_request (GstRTSPClient * client, GstRTSPContext * ctx)
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
{
GstRTSPResult res;
guint8 *data;
guint size;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
g_signal_emit (client,
gst_rtsp_client_signals[SIGNAL_PRE_GET_PARAMETER_REQUEST], 0, ctx,
&sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = gst_rtsp_message_get_body (ctx->request, &data, &size);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
if (res != GST_RTSP_OK)
goto bad_request;
rtsp-client: Also handle the (S|G)ET_PARAMETER case of size==0 || !data as keep-alive If there is no Content-Length header, no body would be allocated and the '\0' would also not be appended to the body.
2017-01-19 12:57:19 +00:00
if (size == 0 || !data || strlen ((char *) data) == 0) {
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (ctx->request->type_data.request.version >= GST_RTSP_VERSION_2_0) {
GST_ERROR_OBJECT (client, "Using PLAY request for keep-alive is forbidden"
" in RTSP 2.0");
goto bad_request;
}
rtsp-client: Also handle the (S|G)ET_PARAMETER case of size==0 || !data as keep-alive If there is no Content-Length header, no body would be allocated and the '\0' would also not be appended to the body.
2017-01-19 12:57:19 +00:00
/* no body (or only '\0'), keep-alive request */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_OK, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
} else {
/* there is a body, handle the params */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = GST_RTSP_CLIENT_GET_CLASS (client)->params_get (client, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
if (res != GST_RTSP_OK)
goto bad_request;
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
}
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_GET_PARAMETER_REQUEST],
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
/* ERRORS */
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
return FALSE;
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
bad_request:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: bad request", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return FALSE;
}
}
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_set_param_request (GstRTSPClient * client, GstRTSPContext * ctx)
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
{
GstRTSPResult res;
guint8 *data;
guint size;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
g_signal_emit (client,
gst_rtsp_client_signals[SIGNAL_PRE_SET_PARAMETER_REQUEST], 0, ctx,
&sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = gst_rtsp_message_get_body (ctx->request, &data, &size);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
if (res != GST_RTSP_OK)
goto bad_request;
rtsp-client: Also handle the (S|G)ET_PARAMETER case of size==0 || !data as keep-alive If there is no Content-Length header, no body would be allocated and the '\0' would also not be appended to the body.
2017-01-19 12:57:19 +00:00
if (size == 0 || !data || strlen ((char *) data) == 0) {
/* no body (or only '\0'), keep-alive request */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_OK, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
} else {
/* there is a body, handle the params */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
res = GST_RTSP_CLIENT_GET_CLASS (client)->params_set (client, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
if (res != GST_RTSP_OK)
goto bad_request;
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
}
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_SET_PARAMETER_REQUEST],
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
/* ERRORS */
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
return FALSE;
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
bad_request:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: bad request", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return FALSE;
}
}
Initial import
2008-10-09 12:29:12 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_pause_request (GstRTSPClient * client, GstRTSPContext * ctx)
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
GstRTSPSession *session;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
GstRTSPClientClass *klass;
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
GstRTSPSessionMedia *sessmedia;
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
GstRTSPMedia *media;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPStatusCode code;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPState rtspstate;
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
gchar *path;
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
gint matched;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
guint i, n;
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!(session = ctx->session))
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
goto no_session;
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!ctx->uri)
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
goto no_uri;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
path = klass->make_path_from_uri (client, ctx->uri);
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* get a handle to the configuration of the media in the session */
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
sessmedia = gst_rtsp_session_dup_media (session, path, &matched);
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
if (!sessmedia)
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
goto not_found;
Initial import
2008-10-09 12:29:12 +00:00
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
if (path[matched] != '\0')
goto no_aggregate;
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
media = gst_rtsp_session_media_get_media (sessmedia);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_object_ref (media);
gst_rtsp_media_lock (media);
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
n = gst_rtsp_media_n_streams (media);
for (i = 0; i < n; i++) {
GstRTSPStream *stream = gst_rtsp_media_get_stream (media, i);
if (gst_rtsp_stream_get_publish_clock_mode (stream) ==
GST_RTSP_PUBLISH_CLOCK_MODE_CLOCK_AND_OFFSET)
goto not_supported;
}
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->sessmedia = sessmedia;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_PAUSE_REQUEST], 0,
ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
rtspstate = gst_rtsp_session_media_get_rtsp_state (sessmedia);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
/* the session state must be playing or recording */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (rtspstate != GST_RTSP_STATE_PLAYING &&
rtspstate != GST_RTSP_STATE_RECORDING)
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
goto invalid_state;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
/* then pause sending */
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
gst_rtsp_session_media_set_state (sessmedia, GST_STATE_PAUSED);
Initial import
2008-10-09 12:29:12 +00:00
/* construct the response now */
code = GST_RTSP_STS_OK;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
Initial import
2008-10-09 12:29:12 +00:00
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
/* the state is now READY */
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
gst_rtsp_session_media_set_rtsp_state (sessmedia, GST_RTSP_STATE_READY);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PAUSE_REQUEST], 0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
Initial import
2008-10-09 12:29:12 +00:00
/* ERRORS */
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_session:
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp-client: Fix typo in debug message
2017-08-14 18:04:23 +00:00
GST_ERROR ("client %p: no session", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
Initial import
2008-10-09 12:29:12 +00:00
return FALSE;
}
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
no_uri:
{
GST_ERROR ("client %p: no uri supplied", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
return FALSE;
}
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
not_found:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no media for uri", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
return FALSE;
}
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
no_aggregate:
{
GST_ERROR ("client %p: no aggregate path %s", client, path);
send_generic_response (client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GST_RTSP_STS_ONLY_AGGREGATE_OPERATION_ALLOWED, ctx);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
return FALSE;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_object_unref (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
return FALSE;
}
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
invalid_state:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: not PLAYING or RECORDING", client);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_VALID_IN_THIS_STATE,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_object_unref (media);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
return FALSE;
}
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
not_supported:
{
GST_ERROR ("client %p: pausing not supported", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_object_unref (media);
rtsp-server: Implement clock signalling according to RFC7273 For NTP and PTP clocks we signal the actual clock that is used and signal the direct media clock offset. For all other clocks we at least signal that it's the local sender clock. This allows receivers to know which clock was used to generate the media and its RTP timestamps. Receivers can then implement network synchronization, either absolute or at least relative by getting the sender clock rate directly via NTP/PTP instead of estimating it from RTP timestamps and packet receive times. https://bugzilla.gnome.org/show_bug.cgi?id=760005
2015-12-30 16:39:05 +00:00
return FALSE;
}
Initial import
2008-10-09 12:29:12 +00:00
}
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
/* convert @url and @path to a URL used as a content base for the factory
* located at @path */
static gchar *
rtsp-server: Fix lots of compiler warnings with clang
2014-02-09 09:45:28 +00:00
make_base_url (GstRTSPClient * client, GstRTSPUrl * url, const gchar * path)
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
{
GstRTSPUrl tmp;
rtsp-server: Fix lots of compiler warnings with clang
2014-02-09 09:45:28 +00:00
gchar *result;
const gchar *trail;
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
/* check for trailing '/' and append one */
trail = (path[strlen (path) - 1] != '/' ? "/" : "");
tmp = *url;
tmp.user = NULL;
tmp.passwd = NULL;
tmp.abspath = g_strdup_printf ("%s%s", path, trail);
tmp.query = NULL;
result = gst_rtsp_url_get_request_uri (&tmp);
g_free (tmp.abspath);
return result;
}
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
/* Check if the given header of type double is present and, if so,
* put it's value in the supplied variable.
*/
static GstRTSPStatusCode
parse_header_value_double (GstRTSPClient * client, GstRTSPContext * ctx,
GstRTSPHeaderField header, gboolean * present, gdouble * value)
{
GstRTSPResult res;
gchar *str;
gchar *end;
res = gst_rtsp_message_get_header (ctx->request, header, &str, 0);
if (res == GST_RTSP_OK) {
*value = g_ascii_strtod (str, &end);
if (end == str)
goto parse_header_failed;
GST_DEBUG ("client %p: got '%s', value %f", client,
gst_rtsp_header_as_text (header), *value);
*present = TRUE;
} else {
*present = FALSE;
}
return GST_RTSP_STS_OK;
parse_header_failed:
{
GST_ERROR ("client %p: failed parsing '%s' header", client,
gst_rtsp_header_as_text (header));
return GST_RTSP_STS_BAD_REQUEST;
}
}
/* Parse scale and speed headers, if present, and set the rate to
* (rate * scale * speed) */
static GstRTSPStatusCode
parse_scale_and_speed (GstRTSPClient * client, GstRTSPContext * ctx,
client: Scale implies TRICK_MODE
2018-11-06 23:33:01 +00:00
gboolean * scale_present, gboolean * speed_present, gdouble * rate,
GstSeekFlags * flags)
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
{
gdouble scale = 1.0;
gdouble speed = 1.0;
GstRTSPStatusCode status;
GST_DEBUG ("got rate %f", *rate);
status = parse_header_value_double (client, ctx, GST_RTSP_HDR_SCALE,
scale_present, &scale);
if (status != GST_RTSP_STS_OK)
return status;
client: compare booleans, not pointers to them
2018-11-06 23:32:29 +00:00
if (*scale_present) {
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
GST_DEBUG ("got Scale %f", scale);
if (scale == 0)
goto bad_scale_value;
*rate *= scale;
client: Scale implies TRICK_MODE
2018-11-06 23:33:01 +00:00
if (ABS (scale) != 1.0)
*flags |= GST_SEEK_FLAG_TRICKMODE;
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
}
GST_DEBUG ("rate after parsing Scale %f", *rate);
status = parse_header_value_double (client, ctx, GST_RTSP_HDR_SPEED,
speed_present, &speed);
if (status != GST_RTSP_STS_OK)
return status;
client: compare booleans, not pointers to them
2018-11-06 23:32:29 +00:00
if (*speed_present) {
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
GST_DEBUG ("got Speed %f", speed);
if (speed <= 0)
goto bad_speed_value;
*rate *= speed;
}
GST_DEBUG ("rate after parsing Speed %f", *rate);
return status;
bad_scale_value:
{
GST_ERROR ("client %p: bad 'Scale' header value (%f)", client, scale);
return GST_RTSP_STS_BAD_REQUEST;
}
bad_speed_value:
{
GST_ERROR ("client %p: bad 'Speed' header value (%f)", client, speed);
return GST_RTSP_STS_BAD_REQUEST;
}
}
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
static GstRTSPStatusCode
setup_play_mode (GstRTSPClient * client, GstRTSPContext * ctx,
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
GstRTSPRangeUnit * unit, gboolean * scale_present, gboolean * speed_present)
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
{
gchar *str;
GstRTSPResult res;
GstRTSPTimeRange *range = NULL;
gdouble rate = 1.0;
GstSeekFlags flags = GST_SEEK_FLAG_NONE;
GstRTSPClientClass *klass = GST_RTSP_CLIENT_GET_CLASS (client);
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
GstRTSPStatusCode rtsp_status_code;
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
GstClockTime trickmode_interval = 0;
gboolean enable_rate_control = TRUE;
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
/* parse the range header if we have one */
res = gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_RANGE, &str, 0);
if (res == GST_RTSP_OK) {
gchar *seek_style = NULL;
res = gst_rtsp_range_parse (str, &range);
if (res != GST_RTSP_OK)
goto parse_range_failed;
*unit = range->unit;
/* parse seek style header, if present */
res = gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_SEEK_STYLE,
&seek_style, 0);
if (res == GST_RTSP_OK) {
if (g_strcmp0 (seek_style, "RAP") == 0)
flags = GST_SEEK_FLAG_ACCURATE;
else if (g_strcmp0 (seek_style, "CoRAP") == 0)
flags = GST_SEEK_FLAG_KEY_UNIT;
else if (g_strcmp0 (seek_style, "First-Prior") == 0)
flags = GST_SEEK_FLAG_KEY_UNIT & GST_SEEK_FLAG_SNAP_BEFORE;
else if (g_strcmp0 (seek_style, "Next") == 0)
flags = GST_SEEK_FLAG_KEY_UNIT & GST_SEEK_FLAG_SNAP_AFTER;
else
GST_FIXME_OBJECT (client, "Add support for seek style %s", seek_style);
rtsp-client: define all seek accuracy flags from setup_play_mode We then pass those to adjust_play_mode, which needs to operate on the "final" seek flags, as previously the code in rtsp-media was assuming that accuracy seek flags (accurate / key_unit) should not be set if the flags passed to the seek method were already set.
2019-07-27 03:14:28 +00:00
} else if (range->min.type == GST_RTSP_TIME_END) {
flags = GST_SEEK_FLAG_ACCURATE;
} else {
flags = GST_SEEK_FLAG_KEY_UNIT;
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
}
rtsp-client: Revitalize dead code Leftover from 65d9aa327cd1844934836249cd4463edf09c725d CID: 1455379
2019-11-21 16:12:45 +00:00
if (seek_style)
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_SEEK_STYLE,
seek_style);
rtsp-client: define all seek accuracy flags from setup_play_mode We then pass those to adjust_play_mode, which needs to operate on the "final" seek flags, as previously the code in rtsp-media was assuming that accuracy seek flags (accurate / key_unit) should not be set if the flags passed to the seek method were already set.
2019-07-27 03:14:28 +00:00
} else {
flags = GST_SEEK_FLAG_ACCURATE;
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
}
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
/* check for scale and/or speed headers
* we will set the seek rate to (speed * scale) and let the media decide
* the resulting scale and speed. in the response we will use rate and applied
* rate from the resulting segment as values for the speed and scale headers
* respectively */
rtsp_status_code = parse_scale_and_speed (client, ctx, scale_present,
client: Scale implies TRICK_MODE
2018-11-06 23:33:01 +00:00
speed_present, &rate, &flags);
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
if (rtsp_status_code != GST_RTSP_STS_OK)
goto scale_speed_failed;
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
/* give the application a chance to tweak range, flags, or rate */
if (klass->adjust_play_mode != NULL) {
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
rtsp_status_code =
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
klass->adjust_play_mode (client, ctx, &range, &flags, &rate,
&trickmode_interval, &enable_rate_control);
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
if (rtsp_status_code != GST_RTSP_STS_OK)
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
goto adjust_play_mode_failed;
}
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
gst_rtsp_media_set_rate_control (ctx->media, enable_rate_control);
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
/* now do the seek with the seek options */
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
gst_rtsp_media_seek_trickmode (ctx->media, range, flags, rate,
trickmode_interval);
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
if (range != NULL)
gst_rtsp_range_free (range);
if (gst_rtsp_media_get_status (ctx->media) == GST_RTSP_MEDIA_STATUS_ERROR)
goto seek_failed;
return GST_RTSP_STS_OK;
parse_range_failed:
{
GST_ERROR ("client %p: failed parsing range header", client);
return GST_RTSP_STS_BAD_REQUEST;
}
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
scale_speed_failed:
{
if (range != NULL)
gst_rtsp_range_free (range);
GST_ERROR ("client %p: failed parsing Scale or Speed headers", client);
return rtsp_status_code;
}
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
adjust_play_mode_failed:
{
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
GST_ERROR ("client %p: sub class returned bad code (%d)", client,
rtsp_status_code);
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
if (range != NULL)
gst_rtsp_range_free (range);
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
return rtsp_status_code;
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
}
seek_failed:
{
GST_ERROR ("client %p: seek failed", client);
return GST_RTSP_STS_SERVICE_UNAVAILABLE;
}
}
Initial import
2008-10-09 12:29:12 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_play_request (GstRTSPClient * client, GstRTSPContext * ctx)
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
GstRTSPSession *session;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
GstRTSPClientClass *klass;
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
GstRTSPSessionMedia *sessmedia;
GstRTSPMedia *media;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPStatusCode code;
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
GstRTSPUrl *uri;
client: store setup uri and use in PLAY response Store the uri used when doing the setup and use that in the PLAY response. fixes https://bugzilla.gnome.org/show_bug.cgi?id=715168
2013-11-28 16:47:18 +00:00
gchar *str;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPState rtspstate;
rtsp-media/client: Reply to PLAY request with same type of Range Remember the type of Range from the PLAY request and use the same type for the reply.
2013-02-22 19:17:29 +00:00
GstRTSPRangeUnit unit = GST_RTSP_RANGE_NPT;
rtsp-client: RTP Info exists conditionally in PLAY If RTP Info is missing and it is not a receiver only, eg. audio backchannel. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR. In rfc2326 it says RTP-info is req. but in RFC7826 it is conditional. Since 1.14 there is audio backchannel support. Thus RTP-info is conditional now. When audio backchannel only mode, there is no RTP-info. Fixes #82
2019-09-25 09:14:08 +00:00
gchar *path, *rtpinfo = NULL;
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
gint matched;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
GPtrArray *transports;
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
gboolean scale_present;
gboolean speed_present;
gdouble rate;
gdouble applied_rate;
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!(session = ctx->session))
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
goto no_session;
Initial import
2008-10-09 12:29:12 +00:00
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
if (!(uri = ctx->uri))
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
goto no_uri;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
path = klass->make_path_from_uri (client, uri);
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* get a handle to the configuration of the media in the session */
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
sessmedia = gst_rtsp_session_dup_media (session, path, &matched);
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
if (!sessmedia)
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
goto not_found;
Initial import
2008-10-09 12:29:12 +00:00
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
if (path[matched] != '\0')
goto no_aggregate;
session-media: let the session-media make the RTPInfo Add method to create the RTPInfo for a stream-transport. Add method to create the RTPInfo for all stream-transports in a session-media. Use the session-media RTPInfo code in client. This allows us to refactor another method to link the TCP callbacks.
2013-12-26 14:41:14 +00:00
g_free (path);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->sessmedia = sessmedia;
ctx->media = media = gst_rtsp_session_media_get_media (sessmedia);
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_object_ref (media);
gst_rtsp_media_lock (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_PLAY_REQUEST], 0,
ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if (!(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_PLAY))
goto unsupported_mode;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
/* the session state must be playing or ready */
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
rtspstate = gst_rtsp_session_media_get_rtsp_state (sessmedia);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (rtspstate != GST_RTSP_STATE_PLAYING && rtspstate != GST_RTSP_STATE_READY)
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
goto invalid_state;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
/* update the pipeline */
transports = gst_rtsp_session_media_get_transports (sessmedia);
if (!gst_rtsp_media_complete_pipeline (media, transports)) {
g_ptr_array_unref (transports);
goto pipeline_error;
}
g_ptr_array_unref (transports);
client: suspend after SDP and unsuspend before PLAYING Based on patches by Ognyan Tonchev <ognyan@axis.com> Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711257
2013-11-28 13:14:35 +00:00
/* in play we first unsuspend, media could be suspended from SDP or PAUSED */
if (!gst_rtsp_media_unsuspend (media))
goto unsuspend_failed;
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
code = setup_play_mode (client, ctx, &unit, &scale_present, &speed_present);
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
if (code != GST_RTSP_STS_OK)
goto invalid_mode;
Add beginnings of seeking. Parse the Range header and perform a seek on the pipeline for the requested position. It's disabled currently until I figure out what's going wrong.
2009-03-12 19:32:14 +00:00
session-media: let the session-media make the RTPInfo Add method to create the RTPInfo for a stream-transport. Add method to create the RTPInfo for all stream-transports in a session-media. Use the session-media RTPInfo code in client. This allows us to refactor another method to link the TCP callbacks.
2013-12-26 14:41:14 +00:00
/* grab RTPInfo from the media now */
rtsp-client: RTP Info when completed_sender Change condition that should be fulfilled regarding RTPInfo. Replace !gst_rtsp_media_is_receive_only with gst_rtsp_media_has_completed_sender. It is more correct to actually look for a sender pipeline that is complete. Only then a RTPInfo should exist. gst_rtsp_media_is_receive_only gives different answears depending on state of server. If Describe is called wth URL+options for backchannel SDP will give only audio and only backchannel a=sendonly If Describe is called on URL+options that gives both audio and video direction from server to client, pipelines are created. Thus receive_only will return false, even though Setup only would setup backchannel. RTP-Info is only for outgoing streams. Thus one should look if outgoing streams are complete.
2019-09-30 13:13:15 +00:00
if (gst_rtsp_media_has_completed_sender (media) &&
rtsp-client: RTP Info exists conditionally in PLAY If RTP Info is missing and it is not a receiver only, eg. audio backchannel. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR. In rfc2326 it says RTP-info is req. but in RFC7826 it is conditional. Since 1.14 there is audio backchannel support. Thus RTP-info is conditional now. When audio backchannel only mode, there is no RTP-info. Fixes #82
2019-09-25 09:14:08 +00:00
!(rtpinfo = gst_rtsp_session_media_get_rtpinfo (sessmedia)))
rtsp-client: RTP Info must exist in PLAY response If RTP Info is missing. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR Fixes #76
2019-08-30 12:00:52 +00:00
goto rtp_info_error;
Initial import
2008-10-09 12:29:12 +00:00
/* construct the response now */
code = GST_RTSP_STS_OK;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
/* add the RTP-Info header */
rtsp-client: RTP Info exists conditionally in PLAY If RTP Info is missing and it is not a receiver only, eg. audio backchannel. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR. In rfc2326 it says RTP-info is req. but in RFC7826 it is conditional. Since 1.14 there is audio backchannel support. Thus RTP-info is conditional now. When audio backchannel only mode, there is no RTP-info. Fixes #82
2019-09-25 09:14:08 +00:00
if (rtpinfo)
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_RTP_INFO,
rtpinfo);
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
/* add the range */
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
str = gst_rtsp_media_get_range_string (media, TRUE, unit);
Do not take range header if range is invalid
2013-07-11 20:11:55 +00:00
if (str)
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_RANGE, str);
Initial import
2008-10-09 12:29:12 +00:00
Use complete streams for scale and speed. Without this patch it's always stream0 that is used to get segment event that is used to set scale and speed. This even if client not doing SETUP for stream0. At least in suspend mode reset this not working since then it's just random if send_rtp_sink have got any segment event. There are no check if send_rtp_sink for stream0 got any data before media is prerolled after PLAY request.
2019-08-21 12:57:25 +00:00
if (gst_rtsp_media_has_completed_sender (media)) {
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
/* the scale and speed headers must always be added if they were present in
* the request. however, even if they were not, we still add them if
* applied_rate or rate deviate from the "normal", i.e. 1.0 */
if (!gst_rtsp_media_get_rates (media, &rate, &applied_rate))
goto get_rates_error;
g_assert (rate != 0 && applied_rate != 0);
if (scale_present || applied_rate != 1.0)
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_SCALE,
g_strdup_printf ("%1.3f", applied_rate));
if (speed_present || rate != 1.0)
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_SPEED,
g_strdup_printf ("%1.3f", rate));
}
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
if (klass->adjust_play_response) {
code = klass->adjust_play_response (client, ctx);
if (code != GST_RTSP_STS_OK)
goto adjust_play_response_failed;
}
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
Initial import
2008-10-09 12:29:12 +00:00
client: fix typo in comment
2014-05-01 03:29:54 +00:00
/* start playing after sending the response */
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
gst_rtsp_session_media_set_state (sessmedia, GST_STATE_PLAYING);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
client: cleanups Rename variables for clarity Keep media in state when we can
2013-07-02 09:58:02 +00:00
gst_rtsp_session_media_set_rtsp_state (sessmedia, GST_RTSP_STATE_PLAYING);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PLAY_REQUEST], 0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
Initial import
2008-10-09 12:29:12 +00:00
/* ERRORS */
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_session:
Initial import
2008-10-09 12:29:12 +00:00
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no session", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
Initial import
2008-10-09 12:29:12 +00:00
return FALSE;
}
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
no_uri:
{
GST_ERROR ("client %p: no uri supplied", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
return FALSE;
}
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
not_found:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: media not found", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
Check if return value of gst_rtsp_session_get_media is not NULL
2009-01-08 15:28:24 +00:00
return FALSE;
}
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
no_aggregate:
{
GST_ERROR ("client %p: no aggregate path %s", client, path);
send_generic_response (client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GST_RTSP_STS_ONLY_AGGREGATE_OPERATION_ALLOWED, ctx);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: append query string in PAUSE/PLAY/TEARDOWN as well
2013-10-04 04:29:30 +00:00
g_free (path);
client: error out on non-aggregate control We require aggregate control (for now) for PLAY, PAUSE and TEARDOWN.
2013-07-03 14:15:04 +00:00
return FALSE;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
return FALSE;
}
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
invalid_state:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: not PLAYING or READY", client);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_VALID_IN_THIS_STATE,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
return FALSE;
}
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
pipeline_error:
{
GST_ERROR ("client %p: failed to configure the pipeline", client);
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_VALID_IN_THIS_STATE,
ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
return FALSE;
}
client: suspend after SDP and unsuspend before PLAYING Based on patches by Ognyan Tonchev <ognyan@axis.com> Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711257
2013-11-28 13:14:35 +00:00
unsuspend_failed:
{
GST_ERROR ("client %p: unsuspend failed", client);
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
client: suspend after SDP and unsuspend before PLAYING Based on patches by Ognyan Tonchev <ognyan@axis.com> Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711257
2013-11-28 13:14:35 +00:00
return FALSE;
}
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
invalid_mode:
rtsp-media: If seeking fails, don't wait forever for the media to preroll again Instead error out properly the same way as if the SEEKING query already failed.
2015-02-12 14:48:46 +00:00
{
GST_ERROR ("client %p: seek failed", client);
rtsp-client: allow sub classes to adjust the seek Adds a new virtual function, adjust_play_mode(), that allows sub classes to adjust the seek done on the media. The sub class can modify the values of the the seek flags and the rate. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-01 16:51:49 +00:00
send_generic_response (client, code, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-media: If seeking fails, don't wait forever for the media to preroll again Instead error out properly the same way as if the SEEKING query already failed.
2015-02-12 14:48:46 +00:00
return FALSE;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
unsupported_mode:
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
{
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
GST_ERROR ("client %p: media does not support PLAY", client);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_ALLOWED, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
get_rates_error:
{
GST_ERROR ("client %p: failed obtaining rate and applied_rate", client);
send_generic_response (client, GST_RTSP_STS_INTERNAL_SERVER_ERROR, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-client: add support for Scale and Speed header Add support for the RTSP Scale and Speed headers by setting the rate in the seek to (scale*speed). We then check the resulting segment for rate and applied rate, and use them as values for the Speed and Scale headers respectively. https://bugzilla.gnome.org/show_bug.cgi?id=754575
2018-10-12 06:53:04 +00:00
return FALSE;
}
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
adjust_play_response_failed:
{
GST_ERROR ("client %p: failed to adjust play response", client);
send_generic_response (client, code, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
onvif: Implement and test the Streaming Specification https://www.onvif.org/specs/stream/ONVIF-Streaming-Spec.pdf
2019-04-04 22:48:07 +00:00
return FALSE;
}
rtsp-client: RTP Info must exist in PLAY response If RTP Info is missing. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR Fixes #76
2019-08-30 12:00:52 +00:00
rtp_info_error:
{
GST_ERROR ("client %p: failed to add RTP-Info", client);
send_generic_response (client, GST_RTSP_STS_INTERNAL_SERVER_ERROR, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
rtsp-client: make sure sessmedia will not get freed while used handle_*_request() functions were all retrieving the session media from the session by calling gst_rtsp_session_get_media () which is a transfer-none call. If a session timeout happens at that time, the session media may get freed making the pointer invalid.. Fixes #757 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1053>
2021-10-03 21:25:23 +00:00
g_object_unref (sessmedia);
rtsp-client: RTP Info must exist in PLAY response If RTP Info is missing. Then return GST_RTSP_STS_INTERNAL_SERVER_ERROR Fixes #76
2019-08-30 12:00:52 +00:00
return FALSE;
}
Initial import
2008-10-09 12:29:12 +00:00
}
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
static void
do_keepalive (GstRTSPSession * session)
{
Use GStreamer's debugging subsystem
2009-11-21 18:20:23 +00:00
GST_INFO ("keep session %p alive", session);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
gst_rtsp_session_touch (session);
}
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
/* parse @transport and return a valid transport in @tr. only transports
client: let stream check supported transport Delegate the check if a transport is allowed to the stream. See https://bugzilla.gnome.org/show_bug.cgi?id=720696
2014-01-07 11:28:47 +00:00
* supported by @stream are returned. Returns FALSE if no valid transport
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
* was found. */
static gboolean
client: let stream check supported transport Delegate the check if a transport is allowed to the stream. See https://bugzilla.gnome.org/show_bug.cgi?id=720696
2014-01-07 11:28:47 +00:00
parse_transport (const char *transport, GstRTSPStream * stream,
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
GstRTSPTransport * tr)
{
gint i;
gboolean res;
gchar **transports;
res = FALSE;
gst_rtsp_transport_init (tr);
rtsp: improve debug
2012-11-15 15:15:20 +00:00
GST_DEBUG ("parsing transports %s", transport);
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
transports = g_strsplit (transport, ",", 0);
/* loop through the transports, try to parse */
for (i = 0; transports[i]; i++) {
client: Strip transport parts as whitespaces could be around commas https://bugzilla.gnome.org/show_bug.cgi?id=758428
2016-01-08 23:12:14 +00:00
g_strstrip (transports[i]);
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
res = gst_rtsp_transport_parse (transports[i], tr);
if (res != GST_RTSP_OK) {
/* no valid transport, search some more */
GST_WARNING ("could not parse transport %s", transports[i]);
goto next;
}
client: let stream check supported transport Delegate the check if a transport is allowed to the stream. See https://bugzilla.gnome.org/show_bug.cgi?id=720696
2014-01-07 11:28:47 +00:00
/* we have a transport, see if it's supported */
if (!gst_rtsp_stream_is_transport_supported (stream, tr)) {
client: refactor transport parsing
2012-10-27 19:26:55 +00:00
GST_WARNING ("unsupported transport %s", transports[i]);
goto next;
}
/* we have a valid transport */
GST_INFO ("found valid transport %s", transports[i]);
res = TRUE;
break;
next:
gst_rtsp_transport_init (tr);
}
g_strfreev (transports);
return res;
}
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
static gboolean
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
default_configure_client_media (GstRTSPClient * client, GstRTSPMedia * media,
GstRTSPStream * stream, GstRTSPContext * ctx)
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
{
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
GstRTSPMessage *request = ctx->request;
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
gchar *blocksize_str;
Get payloader stats only for the sending streams Get/set payloader properties only for streams that actually contain a payloader element. https://bugzilla.gnome.org/show_bug.cgi?id=796523
2018-06-05 06:44:44 +00:00
if (!gst_rtsp_stream_is_sender (stream))
return TRUE;
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
if (gst_rtsp_message_get_header (request, GST_RTSP_HDR_BLOCKSIZE,
client: fix docs
2012-07-10 09:39:58 +00:00
&blocksize_str, 0) == GST_RTSP_OK) {
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
guint64 blocksize;
gchar *end;
blocksize = g_ascii_strtoull (blocksize_str, &end, 10);
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
if (end == blocksize_str)
goto parse_failed;
/* we don't want to change the mtu when this media
* can be shared because it impacts other clients */
if (gst_rtsp_media_is_shared (media))
goto done;
if (blocksize > G_MAXUINT)
blocksize = G_MAXUINT;
gst_rtsp_stream_set_mtu (stream, blocksize);
}
done:
return TRUE;
/* ERRORS */
parse_failed:
{
GST_ERROR_OBJECT (client, "failed to parse blocksize");
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
return FALSE;
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
}
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
static gboolean
rtsp-client: Make configure_client_transport virtual This patch makes configure_client_transport virtual. The functionality is needed to handle some weird clients sending multicast transport settings as url options. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702173
2013-06-17 14:18:37 +00:00
default_configure_client_transport (GstRTSPClient * client,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext * ctx, GstRTSPTransport * ct)
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* we have a valid transport now, set the destination of the client. */
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
if (ct->lower_transport == GST_RTSP_LOWER_TRANS_UDP_MCAST ||
ct->lower_transport == GST_RTSP_LOWER_TRANS_UDP) {
/* allocate UDP ports */
GSocketFamily family;
gboolean use_client_settings = FALSE;
family = priv->is_ipv6 ? G_SOCKET_FAMILY_IPV6 : G_SOCKET_FAMILY_IPV4;
rtsp-client: always allocate both IPV4 and IPV6 sockets multiudpsink does not support setting the socket* properties after it has started, which meant that rtsp-server could no longer serve on both IPV4 and IPV6 sockets since the patches from https://bugzilla.gnome.org/show_bug.cgi?id=757488 were merged. When first connecting an IPV6 client then an IPV4 client, multiudpsink fell back to using the IPV6 socket. When first connecting an IPV4 client, then an IPV6 client, multiudpsink errored out, released the IPV4 socket, then crashed when trying to send a message on NULL nevertheless, that is however a separate issue. This could probably be fixed by handling the setting of sockets in multiudpsink after it has started, that will however be a much more significant effort. For now, this commit simply partially reverts the behaviour of rtsp-stream: it will continue to only create the udpsinks when needed, as was the case since the patches were merged, it will however when creating them, always allocate both sockets and set them on the sink before it starts, as was the case prior to the patches. Transport configuration will only error out if the allocation of UDP sockets fails for the actual client's family, this also downgrades the GST_ERRORs in alloc_ports_one_family to GST_WARNINGs, as failing to allocate is no longer necessarily fatal. https://bugzilla.gnome.org/show_bug.cgi?id=796875
2018-07-25 17:54:55 +00:00
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
if ((ct->lower_transport == GST_RTSP_LOWER_TRANS_UDP_MCAST) &&
gst_rtsp_auth_check (GST_RTSP_AUTH_CHECK_TRANSPORT_CLIENT_SETTINGS) &&
stream: Choose the maximum ttl value provided by multicast clients The maximum ttl value provided so far by the multicast clients will be chosen and reported in the response to the current client request. Change-Id: I5408646e3b5a0a224d907ae215bdea60c4f1905f https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-25 13:33:18 +00:00
(ct->destination != NULL)) {
if (!gst_rtsp_stream_verify_mcast_ttl (ctx->stream, ct->ttl))
goto error_ttl;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
use_client_settings = TRUE;
stream: Choose the maximum ttl value provided by multicast clients The maximum ttl value provided so far by the multicast clients will be chosen and reported in the response to the current client request. Change-Id: I5408646e3b5a0a224d907ae215bdea60c4f1905f https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-25 13:33:18 +00:00
}
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
rtsp-client: always allocate both IPV4 and IPV6 sockets multiudpsink does not support setting the socket* properties after it has started, which meant that rtsp-server could no longer serve on both IPV4 and IPV6 sockets since the patches from https://bugzilla.gnome.org/show_bug.cgi?id=757488 were merged. When first connecting an IPV6 client then an IPV4 client, multiudpsink fell back to using the IPV6 socket. When first connecting an IPV4 client, then an IPV6 client, multiudpsink errored out, released the IPV4 socket, then crashed when trying to send a message on NULL nevertheless, that is however a separate issue. This could probably be fixed by handling the setting of sockets in multiudpsink after it has started, that will however be a much more significant effort. For now, this commit simply partially reverts the behaviour of rtsp-stream: it will continue to only create the udpsinks when needed, as was the case since the patches were merged, it will however when creating them, always allocate both sockets and set them on the sink before it starts, as was the case prior to the patches. Transport configuration will only error out if the allocation of UDP sockets fails for the actual client's family, this also downgrades the GST_ERRORs in alloc_ports_one_family to GST_WARNINGs, as failing to allocate is no longer necessarily fatal. https://bugzilla.gnome.org/show_bug.cgi?id=796875
2018-07-25 17:54:55 +00:00
/* We need to allocate the sockets for both families before starting
* multiudpsink, otherwise multiudpsink won't accept new clients with
* a different family.
*/
/* FIXME: could be more adequately solved by making it possible
* to set a socket on multiudpsink after it has already been started */
Fix indentation again
2018-08-14 11:31:55 +00:00
if (!gst_rtsp_stream_allocate_udp_sockets (ctx->stream,
G_SOCKET_FAMILY_IPV4, ct, use_client_settings)
&& family == G_SOCKET_FAMILY_IPV4)
rtsp-client: always allocate both IPV4 and IPV6 sockets multiudpsink does not support setting the socket* properties after it has started, which meant that rtsp-server could no longer serve on both IPV4 and IPV6 sockets since the patches from https://bugzilla.gnome.org/show_bug.cgi?id=757488 were merged. When first connecting an IPV6 client then an IPV4 client, multiudpsink fell back to using the IPV6 socket. When first connecting an IPV4 client, then an IPV6 client, multiudpsink errored out, released the IPV4 socket, then crashed when trying to send a message on NULL nevertheless, that is however a separate issue. This could probably be fixed by handling the setting of sockets in multiudpsink after it has started, that will however be a much more significant effort. For now, this commit simply partially reverts the behaviour of rtsp-stream: it will continue to only create the udpsinks when needed, as was the case since the patches were merged, it will however when creating them, always allocate both sockets and set them on the sink before it starts, as was the case prior to the patches. Transport configuration will only error out if the allocation of UDP sockets fails for the actual client's family, this also downgrades the GST_ERRORs in alloc_ports_one_family to GST_WARNINGs, as failing to allocate is no longer necessarily fatal. https://bugzilla.gnome.org/show_bug.cgi?id=796875
2018-07-25 17:54:55 +00:00
goto error_allocating_ports;
Fix indentation again
2018-08-14 11:31:55 +00:00
if (!gst_rtsp_stream_allocate_udp_sockets (ctx->stream,
G_SOCKET_FAMILY_IPV6, ct, use_client_settings)
&& family == G_SOCKET_FAMILY_IPV6)
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
goto error_allocating_ports;
if (ct->lower_transport == GST_RTSP_LOWER_TRANS_UDP_MCAST) {
stream: Choose the maximum ttl value provided by multicast clients The maximum ttl value provided so far by the multicast clients will be chosen and reported in the response to the current client request. Change-Id: I5408646e3b5a0a224d907ae215bdea60c4f1905f https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-25 13:33:18 +00:00
if (use_client_settings) {
/* FIXME: the address has been successfully allocated, however, in
* the use_client_settings case we need to verify that the allocated
* address is the one requested by the client and if this address is
* an allowed destination. Verifying this via the address pool in not
* the proper way as the address pool should only be used for choosing
* the server-selected address/port pairs. */
GSocket *rtp_socket;
guint ttl;
rtp_socket =
gst_rtsp_stream_get_rtp_multicast_socket (ctx->stream, family);
if (rtp_socket == NULL)
goto no_socket;
ttl = g_socket_get_multicast_ttl (rtp_socket);
g_object_unref (rtp_socket);
if (ct->ttl < ttl) {
/* use the maximum ttl that is requested by multicast clients */
GST_DEBUG ("requested ttl %u, but keeping ttl %u", ct->ttl, ttl);
ct->ttl = ttl;
}
} else {
client: Don't reserve multicast address in the client setting case When two multicast clients request specific transport configurations, and "transport.client-settings" parameter is set to true, it's wrong to actually require that these two clients request the same multicast group. Removed test_client_multicast_invalid_transport_specific test cases as they wrongly require that the requested destination address is supposed to be present in the address pool, also in the case when "transport.client-settings" parameter is set to true. Change-Id: I4580182ef35996caf644686d6139f72ec599c9fa https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-24 12:02:40 +00:00
GstRTSPAddress *addr = NULL;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
g_free (ct->destination);
addr = gst_rtsp_stream_get_multicast_address (ctx->stream, family);
if (addr == NULL)
goto no_address;
ct->destination = g_strdup (addr->address);
ct->port.min = addr->port;
ct->port.max = addr->port + addr->n_ports - 1;
ct->ttl = addr->ttl;
client: Don't reserve multicast address in the client setting case When two multicast clients request specific transport configurations, and "transport.client-settings" parameter is set to true, it's wrong to actually require that these two clients request the same multicast group. Removed test_client_multicast_invalid_transport_specific test cases as they wrongly require that the requested destination address is supposed to be present in the address pool, also in the case when "transport.client-settings" parameter is set to true. Change-Id: I4580182ef35996caf644686d6139f72ec599c9fa https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-24 12:02:40 +00:00
gst_rtsp_address_free (addr);
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
}
stream: Added a list of multicast client addresses When media is shared, the same media stream can be sent to multiple multicast groups. Currently, there is no API to retrieve multicast addresses from the stream. When calling gst_rtsp_stream_get_multicast_address() function, only the first multicast address is returned. With this patch, each multicast destination requested in SETUP will be stored in an internal list (call to gst_rtsp_stream_add_multicast_client_address()). The list of multicast groups requested by the clients can be retrieved by calling gst_rtsp_stream_get_multicast_client_addresses(). There still exist some problems with the current implementation in the multicast case: 1) The receiving part is currently only configured with regard to the first multicast client (see https://bugzilla.gnome.org/show_bug.cgi?id=796917). 2) Secondly, of security reasons, some constraints should be put on the requested multicast destinations (see https://bugzilla.gnome.org/show_bug.cgi?id=796916). Change-Id: I6b060746e472a0734cc2fd828ffe4ea2956733ea https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-26 10:01:16 +00:00
if (!gst_rtsp_stream_add_multicast_client_address (ctx->stream,
ct->destination, ct->port.min, ct->port.max, family))
goto error_mcast_transport;
client: Check client provided addresses against the address pool
2013-02-13 23:32:20 +00:00
} else {
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
GstRTSPUrl *url;
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
url = gst_rtsp_connection_get_url (priv->connection);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
g_free (ct->destination);
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
ct->destination = g_strdup (url->host);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
}
} else {
GstRTSPUrl *url;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
url = gst_rtsp_connection_get_url (priv->connection);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
g_free (ct->destination);
ct->destination = g_strdup (url->host);
if (ct->lower_transport & GST_RTSP_LOWER_TRANS_TCP) {
client: store TCP ports in transport Store the TCP ports in the transport when we are doing RTSP over TCP. This way, we can easily get to the ports from the transport. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=729776
2014-05-20 13:57:30 +00:00
GSocket *sock;
GSocketAddress *addr;
sock = gst_rtsp_connection_get_read_socket (priv->connection);
if ((addr = g_socket_get_remote_address (sock, NULL))) {
/* our read port is the sender port of client */
ct->client_port.min =
g_inet_socket_address_get_port (G_INET_SOCKET_ADDRESS (addr));
g_object_unref (addr);
}
if ((addr = g_socket_get_local_address (sock, NULL))) {
ct->server_port.max =
g_inet_socket_address_get_port (G_INET_SOCKET_ADDRESS (addr));
g_object_unref (addr);
}
sock = gst_rtsp_connection_get_write_socket (priv->connection);
if ((addr = g_socket_get_remote_address (sock, NULL))) {
/* our write port is the receiver port of client */
ct->client_port.max =
g_inet_socket_address_get_port (G_INET_SOCKET_ADDRESS (addr));
g_object_unref (addr);
}
if ((addr = g_socket_get_local_address (sock, NULL))) {
ct->server_port.min =
g_inet_socket_address_get_port (G_INET_SOCKET_ADDRESS (addr));
g_object_unref (addr);
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* check if the client selected channels for TCP */
if (ct->interleaved.min == -1 || ct->interleaved.max == -1) {
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_session_media_alloc_channels (ctx->sessmedia,
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
&ct->interleaved);
}
rtsp-client: Avoid reuse of channel numbers for interleaved If a (strange) client would reuse interleaved channel numbers in multiple SETUP requests, we should not accept them. The channel numbers are used for looking up stream transports in the priv->transports hash table, and transports disappear from the table if channel numbers are reused. RFC 7826 (RTSP 2.0), Section 18.54, clarifies that it is OK for the server to change the channel numbers suggested by the client. https://bugzilla.gnome.org/show_bug.cgi?id=796988
2018-08-17 07:54:27 +00:00
/* alloc new channels if they are already taken */
while (g_hash_table_contains (priv->transports,
GINT_TO_POINTER (ct->interleaved.min))
|| g_hash_table_contains (priv->transports,
GINT_TO_POINTER (ct->interleaved.max))) {
gst_rtsp_session_media_alloc_channels (ctx->sessmedia,
&ct->interleaved);
if (ct->interleaved.max > 255)
goto error_allocating_channels;
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
}
}
return TRUE;
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
/* ERRORS */
stream: Choose the maximum ttl value provided by multicast clients The maximum ttl value provided so far by the multicast clients will be chosen and reported in the response to the current client request. Change-Id: I5408646e3b5a0a224d907ae215bdea60c4f1905f https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-25 13:33:18 +00:00
error_ttl:
{
GST_ERROR_OBJECT (client,
"Failed to allocate UDP ports: invalid ttl value");
return FALSE;
}
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
error_allocating_ports:
{
GST_ERROR_OBJECT (client, "Failed to allocate UDP ports");
return FALSE;
}
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
no_address:
{
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
GST_ERROR_OBJECT (client, "Failed to acquire address for stream");
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
return FALSE;
}
stream: Choose the maximum ttl value provided by multicast clients The maximum ttl value provided so far by the multicast clients will be chosen and reported in the response to the current client request. Change-Id: I5408646e3b5a0a224d907ae215bdea60c4f1905f https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-25 13:33:18 +00:00
no_socket:
{
GST_ERROR_OBJECT (client, "Failed to get UDP socket");
return FALSE;
}
stream: Added a list of multicast client addresses When media is shared, the same media stream can be sent to multiple multicast groups. Currently, there is no API to retrieve multicast addresses from the stream. When calling gst_rtsp_stream_get_multicast_address() function, only the first multicast address is returned. With this patch, each multicast destination requested in SETUP will be stored in an internal list (call to gst_rtsp_stream_add_multicast_client_address()). The list of multicast groups requested by the clients can be retrieved by calling gst_rtsp_stream_get_multicast_client_addresses(). There still exist some problems with the current implementation in the multicast case: 1) The receiving part is currently only configured with regard to the first multicast client (see https://bugzilla.gnome.org/show_bug.cgi?id=796917). 2) Secondly, of security reasons, some constraints should be put on the requested multicast destinations (see https://bugzilla.gnome.org/show_bug.cgi?id=796916). Change-Id: I6b060746e472a0734cc2fd828ffe4ea2956733ea https://bugzilla.gnome.org/show_bug.cgi?id=793441
2018-07-26 10:01:16 +00:00
error_mcast_transport:
{
GST_ERROR_OBJECT (client, "Failed to add multicast client transport");
return FALSE;
}
rtsp-client: Avoid reuse of channel numbers for interleaved If a (strange) client would reuse interleaved channel numbers in multiple SETUP requests, we should not accept them. The channel numbers are used for looking up stream transports in the priv->transports hash table, and transports disappear from the table if channel numbers are reused. RFC 7826 (RTSP 2.0), Section 18.54, clarifies that it is OK for the server to change the channel numbers suggested by the client. https://bugzilla.gnome.org/show_bug.cgi?id=796988
2018-08-17 07:54:27 +00:00
error_allocating_channels:
{
GST_ERROR_OBJECT (client, "Failed to allocate interleaved channels");
return FALSE;
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
}
static GstRTSPTransport *
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
make_server_transport (GstRTSPClient * client, GstRTSPMedia * media,
GstRTSPContext * ctx, GstRTSPTransport * ct)
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
{
GstRTSPTransport *st;
rtspstream: handle both ipv4 and ipv6 clients Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701129
2013-05-30 06:07:48 +00:00
GInetAddress *addr;
GSocketFamily family;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* prepare the server transport */
gst_rtsp_transport_new (&st);
st->trans = ct->trans;
st->profile = ct->profile;
st->lower_transport = ct->lower_transport;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
st->mode_play = ct->mode_play;
st->mode_record = ct->mode_record;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
rtspstream: handle both ipv4 and ipv6 clients Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701129
2013-05-30 06:07:48 +00:00
addr = g_inet_address_new_from_string (ct->destination);
if (!addr) {
GST_ERROR ("failed to get inet addr from client destination");
family = G_SOCKET_FAMILY_IPV4;
} else {
family = g_inet_address_get_family (addr);
g_object_unref (addr);
addr = NULL;
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
switch (st->lower_transport) {
case GST_RTSP_LOWER_TRANS_UDP:
st->client_port = ct->client_port;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_stream_get_server_port (ctx->stream, &st->server_port, family);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
break;
case GST_RTSP_LOWER_TRANS_UDP_MCAST:
st->port = ct->port;
st->destination = g_strdup (ct->destination);
st->ttl = ct->ttl;
break;
case GST_RTSP_LOWER_TRANS_TCP:
st->interleaved = ct->interleaved;
client: store TCP ports in transport Store the TCP ports in the transport when we are doing RTSP over TCP. This way, we can easily get to the ports from the transport. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=729776
2014-05-20 13:57:30 +00:00
st->client_port = ct->client_port;
st->server_port = ct->server_port;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
default:
break;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if ((gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_PLAY))
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
gst_rtsp_stream_get_ssrc (ctx->stream, &st->ssrc);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
return st;
}
rtsp-client: Fix some race conditions around timeout source removal We always need to take the lock while accessing it as otherwise another thread might've removed it in the meantime. Also when destroying and creating a new one, ensure that the mutex is not shortly unlocked in between as during that time another one might potentially be created already. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/119>
2020-05-14 07:08:32 +00:00
static void
rtsp_ctrl_timeout_remove_unlocked (GstRTSPClientPrivate * priv)
{
if (priv->rtsp_ctrl_timeout != NULL) {
GST_DEBUG ("rtsp control session removed timeout %p.",
priv->rtsp_ctrl_timeout);
g_source_destroy (priv->rtsp_ctrl_timeout);
g_source_unref (priv->rtsp_ctrl_timeout);
priv->rtsp_ctrl_timeout = NULL;
priv->rtsp_ctrl_timeout_cnt = 0;
}
}
static void
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove (GstRTSPClient * client)
rtsp-client: Fix some race conditions around timeout source removal We always need to take the lock while accessing it as otherwise another thread might've removed it in the meantime. Also when destroying and creating a new one, ensure that the mutex is not shortly unlocked in between as during that time another one might potentially be created already. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/119>
2020-05-14 07:08:32 +00:00
{
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
g_mutex_lock (&client->priv->lock);
rtsp_ctrl_timeout_remove_unlocked (client->priv);
g_mutex_unlock (&client->priv->lock);
}
static void
rtsp_ctrl_timeout_destroy_notify (gpointer user_data)
{
GWeakRef *client_weak_ref = (GWeakRef *) user_data;
g_weak_ref_clear (client_weak_ref);
g_free (client_weak_ref);
rtsp-client: Fix some race conditions around timeout source removal We always need to take the lock while accessing it as otherwise another thread might've removed it in the meantime. Also when destroying and creating a new one, ensure that the mutex is not shortly unlocked in between as during that time another one might potentially be created already. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/119>
2020-05-14 07:08:32 +00:00
}
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
static gboolean
rtsp_ctrl_timeout_cb (gpointer user_data)
{
gboolean res = G_SOURCE_CONTINUE;
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
GstRTSPClientPrivate *priv;
GWeakRef *client_weak_ref = (GWeakRef *) user_data;
GstRTSPClient *client = (GstRTSPClient *) g_weak_ref_get (client_weak_ref);
if (client == NULL) {
return G_SOURCE_REMOVE;
}
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
priv = client->priv;
g_mutex_lock (&priv->lock);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
priv->rtsp_ctrl_timeout_cnt += RTSP_CTRL_CB_INTERVAL;
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
if ((priv->rtsp_ctrl_timeout_cnt > RTSP_CTRL_TIMEOUT_VALUE)
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
|| (priv->had_session
&& priv->rtsp_ctrl_timeout_cnt > priv->post_session_timeout)) {
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
GST_DEBUG ("rtsp control session timeout %p expired, closing client.",
priv->rtsp_ctrl_timeout);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove_unlocked (client->priv);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
res = G_SOURCE_REMOVE;
}
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
g_mutex_unlock (&priv->lock);
if (res == G_SOURCE_REMOVE) {
gst_rtsp_client_close (client);
}
g_object_unref (client);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
return res;
}
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
static gchar *
gst: Run everything through gst-indent again
2018-04-04 07:06:06 +00:00
stream_make_keymgmt (GstRTSPClient * client, const gchar * location,
GstRTSPStream * stream)
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
{
gchar *base64, *result = NULL;
GstMIKEYMessage *mikey_msg;
GstCaps *srtcpparams;
GstElement *rtcp_encoder;
gint srtcp_cipher, srtp_cipher;
gint srtcp_auth, srtp_auth;
GstBuffer *key;
GType ciphertype, authtype;
GEnumClass *cipher_enum, *auth_enum;
gst: Run everything through gst-indent again
2018-04-04 07:06:06 +00:00
GEnumValue *srtcp_cipher_value, *srtp_cipher_value, *srtcp_auth_value,
*srtp_auth_value;
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
rtcp_encoder = gst_rtsp_stream_get_srtp_encoder (stream);
if (!rtcp_encoder)
goto done;
ciphertype = g_type_from_name ("GstSrtpCipherType");
authtype = g_type_from_name ("GstSrtpAuthType");
cipher_enum = g_type_class_ref (ciphertype);
auth_enum = g_type_class_ref (authtype);
/* We need to bring the encoder to READY so that it generates its key */
gst_element_set_state (rtcp_encoder, GST_STATE_READY);
gst: Run everything through gst-indent again
2018-04-04 07:06:06 +00:00
g_object_get (rtcp_encoder, "rtcp-cipher", &srtcp_cipher, "rtcp-auth",
&srtcp_auth, "rtp-cipher", &srtp_cipher, "rtp-auth", &srtp_auth, "key",
&key, NULL);
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
g_object_unref (rtcp_encoder);
srtcp_cipher_value = g_enum_get_value (cipher_enum, srtcp_cipher);
srtp_cipher_value = g_enum_get_value (cipher_enum, srtp_cipher);
srtcp_auth_value = g_enum_get_value (auth_enum, srtcp_auth);
srtp_auth_value = g_enum_get_value (auth_enum, srtp_auth);
g_type_class_unref (cipher_enum);
g_type_class_unref (auth_enum);
srtcpparams = gst_caps_new_simple ("application/x-srtcp",
"srtcp-cipher", G_TYPE_STRING, srtcp_cipher_value->value_nick,
"srtcp-auth", G_TYPE_STRING, srtcp_auth_value->value_nick,
"srtp-cipher", G_TYPE_STRING, srtp_cipher_value->value_nick,
"srtp-auth", G_TYPE_STRING, srtp_auth_value->value_nick,
"srtp-key", GST_TYPE_BUFFER, key, NULL);
mikey_msg = gst_mikey_message_new_from_caps (srtcpparams);
if (mikey_msg) {
guint send_ssrc;
gst_rtsp_stream_get_ssrc (stream, &send_ssrc);
gst_mikey_message_add_cs_srtp (mikey_msg, 0, send_ssrc, 0);
base64 = gst_mikey_message_base64_encode (mikey_msg);
gst_mikey_message_unref (mikey_msg);
if (base64) {
result = gst_sdp_make_keymgmt (location, base64);
g_free (base64);
}
}
done:
return result;
}
Initial import
2008-10-09 12:29:12 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_setup_request (GstRTSPClient * client, GstRTSPContext * ctx)
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPResult res;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
GstRTSPUrl *uri;
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
gchar *transport, *keymgmt;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPTransport *ct, *st;
GstRTSPStatusCode code;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
GstRTSPSession *session;
rtsp: massive refactoring Make GObjects from the remaining simple structures. Remove GstRTSPSessionStream, it's not needed. Rename GstRTSPMediaStream -> GstRTSPStream: It is shorter Rename GstRTSPMediaTrans -> GstRTSPStreamTransport: It describes how a GstRTSPStream should be transported to a client. Rename GstRTSPMediaFactory::get_element -> create_element because that more accurately describes what it does. Make nice methods instead of poking in the structures. Move some methods inside the relevant object source code. Use GPtrArray to store objects instead of plain arrays, it is more natural and allows us to more easily clean up. Move the allocation of udp ports to the Stream object. The Stream object contains the elements needed to stream the media to a client. Improve the prepare and unprepare methods. Unprepare should now undo everything prepare did. Improve also async unprepare when doing EOS on shutdown. Make sure we always unprepare correctly.
2012-10-25 19:29:58 +00:00
GstRTSPStreamTransport *trans;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
gchar *trans_str;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
GstRTSPSessionMedia *sessmedia;
GstRTSPMedia *media;
GstRTSPStream *stream;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPState rtspstate;
rtsp-client: Make configure_client_transport virtual This patch makes configure_client_transport virtual. The functionality is needed to handle some weird clients sending multicast transport settings as url options. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702173
2013-06-17 14:18:37 +00:00
GstRTSPClientClass *klass;
rtsp-client: If we have a single-stream media and SETUP contains no control, use the one and only stream
2015-01-19 12:09:20 +00:00
gchar *path, *control = NULL;
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
gint matched;
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
gboolean new_session = FALSE;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
gchar *pipelined_request_id = NULL, *accept_range = NULL;
Initial import
2008-10-09 12:29:12 +00:00
Revert "client: map URL to path in requests" This reverts commit e3fded2cec897a2ec003450607b916cc1601fd2d. This is not correct, we only remap the URL to a path in DESCRIBE, the SDP then contains the base and control urls which are used in the SETUP, PLAY, PAUSE and TEARDOWN requests.
2013-09-17 09:41:57 +00:00
if (!ctx->uri)
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
goto no_uri;
Revert "client: map URL to path in requests" This reverts commit e3fded2cec897a2ec003450607b916cc1601fd2d. This is not correct, we only remap the URL to a path in DESCRIBE, the SDP then contains the base and control urls which are used in the SETUP, PLAY, PAUSE and TEARDOWN requests.
2013-09-17 09:41:57 +00:00
uri = ctx->uri;
client: make make_path_from_uri a vmethod
2013-11-07 12:22:09 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
path = klass->make_path_from_uri (client, uri);
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* parse the transport */
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
res =
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_TRANSPORT,
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
&transport, 0);
Initial import
2008-10-09 12:29:12 +00:00
if (res != GST_RTSP_OK)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
goto no_transport;
Initial import
2008-10-09 12:29:12 +00:00
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
/* Handle Pipelined-requests if using >= 2.0 */
if (ctx->request->type_data.request.version >= GST_RTSP_VERSION_2_0)
gst_rtsp_message_get_header (ctx->request,
GST_RTSP_HDR_PIPELINED_REQUESTS, &pipelined_request_id, 0);
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* we create the session after parsing stuff so that we don't make
* a session for malformed requests */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (priv->session_pool == NULL)
Cleanups and reuse media from DESCRIBE Handle thread create errors. Rename some internal methods to better match what they actually do. Handle misconfiguration of session_pool and media_mapping gracefully. Cache the DESCRIBE media and uri in the client connection and reuse them when we receive a SETUP request in the same connection for the same uri. Cleanup the client connection object.
2009-01-29 17:51:02 +00:00
goto no_pool;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
session = ctx->session;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
Add support for session keepalive Get and update the session timeout for all requests. get the session as early as possible.
2009-02-13 18:52:05 +00:00
if (session) {
g_object_ref (session);
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
/* get a handle to the configuration of the media in the session, this can
* return NULL if this is a new url to manage in this session. */
session: use path matching for session media Use a path string instead of a uri to lookup session media in the sessions. Also use path matching to find the largest possible path that matches.
2013-07-03 10:37:48 +00:00
sessmedia = gst_rtsp_session_get_media (session, path, &matched);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
} else {
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
/* we need a new media configuration in this session */
sessmedia = NULL;
}
/* we have no session media, find one and manage it */
if (sessmedia == NULL) {
/* get a handle to the configuration of the media in the session */
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
media = find_media (client, ctx, path, &matched);
rtsp-client: suspend media during setup request SETUP request from clients needs to suspend the media to clear the prerolled buffers. Otherwise it will not affect the prerolled buffer and the prerolled buffers will be incorrect (for example block-size from setup request will not affect the prerolled buffer unless the media is suspended). https://bugzilla.gnome.org/show_bug.cgi?id=758268
2015-11-18 10:14:39 +00:00
/* need to suspend the media, if the protocol has changed */
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
if (media != NULL) {
gst_rtsp_media_lock (media);
rtsp-client: suspend media during setup request SETUP request from clients needs to suspend the media to clear the prerolled buffers. Otherwise it will not affect the prerolled buffer and the prerolled buffers will be incorrect (for example block-size from setup request will not affect the prerolled buffer unless the media is suspended). https://bugzilla.gnome.org/show_bug.cgi?id=758268
2015-11-18 10:14:39 +00:00
gst_rtsp_media_suspend (media);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
} else {
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
if ((media = gst_rtsp_session_media_get_media (sessmedia))) {
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
g_object_ref (media);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_lock (media);
} else {
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
goto media_not_found;
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
}
/* no media, not found then */
if (media == NULL)
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
goto media_not_found_no_reply;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
rtsp-client: If we have a single-stream media and SETUP contains no control, use the one and only stream
2015-01-19 12:09:20 +00:00
if (path[matched] == '\0') {
if (gst_rtsp_media_n_streams (media) == 1) {
stream = gst_rtsp_media_get_stream (media, 0);
} else {
goto control_not_found;
}
} else {
/* path is what matched. */
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
gchar *newpath = g_strndup (path, matched);
rtsp-client: If we have a single-stream media and SETUP contains no control, use the one and only stream
2015-01-19 12:09:20 +00:00
/* control is remainder */
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
if (matched == 1 && path[0] == '/')
control = g_strdup (&path[1]);
else
control = g_strdup (&path[matched + 1]);
g_free (path);
path = newpath;
client: Do not read beyond end of path string If the setup was done without a control url, make sure we don't try to read the non-existing control string and crash.
2013-09-19 13:44:26 +00:00
rtsp-client: If we have a single-stream media and SETUP contains no control, use the one and only stream
2015-01-19 12:09:20 +00:00
/* find the stream now using the control part */
stream = gst_rtsp_media_find_stream (media, control);
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (stream == NULL)
goto stream_not_found;
/* now we have a uri identifying a valid media and stream */
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->stream = stream;
ctx->media = media;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_SETUP_REQUEST], 0,
ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
if (session == NULL) {
Initial import
2008-10-09 12:29:12 +00:00
/* create a session if this fails we probably reached our session limit or
* something. */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (!(session = gst_rtsp_session_pool_create (priv->session_pool)))
Initial import
2008-10-09 12:29:12 +00:00
goto service_unavailable;
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
/* Pipelined requests should be cleared between sessions */
g_hash_table_remove_all (priv->pipelined_requests);
client: also watch newly created session When we newly created a session, start watching it immediately instead of on the next request.
2013-06-20 10:20:21 +00:00
/* make sure this client is closed when the session is closed */
client_watch_session (client, session);
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
new_session = TRUE;
client: emit new-session when new session is created Only emit new-session when we created a new session for a client, not when a client picked up a previous session. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701587
2013-06-20 10:16:07 +00:00
/* signal new session */
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_NEW_SESSION], 0,
session);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->session = session;
Initial import
2008-10-09 12:29:12 +00:00
}
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (pipelined_request_id) {
g_hash_table_insert (client->priv->pipelined_requests,
g_strdup (pipelined_request_id),
g_strdup (gst_rtsp_session_get_sessionid (session)));
}
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
/* Remember that we had at least one session in the past */
priv->had_session = TRUE;
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove (client);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
if (!klass->configure_client_media (client, media, stream, ctx))
goto configure_media_failed_no_reply;
rtsp: allocate channels in TCP mode When the client does not provide us with channels in TCP mode, allocate channels ourselves.
2009-07-27 17:42:44 +00:00
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
gst_rtsp_transport_new (&ct);
/* parse and find a usable supported transport */
client: let stream check supported transport Delegate the check if a transport is allowed to the stream. See https://bugzilla.gnome.org/show_bug.cgi?id=720696
2014-01-07 11:28:47 +00:00
if (!parse_transport (transport, stream, ct))
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
goto unsupported_transports;
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if ((ct->mode_play
&& !(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_PLAY)) || (ct->mode_record
&& !(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_RECORD)))
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
goto unsupported_mode;
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
/* parse the keymgmt */
if (gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_KEYMGMT,
&keymgmt, 0) == GST_RTSP_OK) {
rtsp-stream: extract handle_keymgmt from rtsp-client rtspclientsink will also need to parse KeyMgmt headers sent by the server to decrypt the RTCP backchannel stream https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:06:31 +00:00
if (!gst_rtsp_stream_handle_keymgmt (ctx->stream, keymgmt))
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
goto keymgmt_error;
}
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_ACCEPT_RANGES,
&accept_range, 0) == GST_RTSP_OK) {
GEnumValue *runit = NULL;
gint i;
gchar **valid_ranges;
GEnumClass *runit_class = g_type_class_ref (GST_TYPE_RTSP_RANGE_UNIT);
gst_rtsp_message_dump (ctx->request);
valid_ranges = g_strsplit (accept_range, ",", -1);
for (i = 0; valid_ranges[i]; i++) {
gchar *range = valid_ranges[i];
while (*range == ' ')
range++;
runit = g_enum_get_value_by_nick (runit_class, range);
if (runit)
break;
}
g_strfreev (valid_ranges);
g_type_class_unref (runit_class);
if (!runit)
goto unsupported_range_unit;
}
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
if (sessmedia == NULL) {
/* manage the media in our session now, if not done already */
rtsp-client: Fix leaking of media in error cases With additional fixes by Kseniya Vasilchuk <vasilchukkseniia@gmail.com> and myself to make the media refcounting a bit easier to follow. https://bugzilla.gnome.org/show_bug.cgi?id=755632
2015-09-25 15:04:00 +00:00
sessmedia =
gst_rtsp_session_manage_media (session, path, g_object_ref (media));
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
/* if we stil have no media, error */
if (sessmedia == NULL)
goto sessmedia_unavailable;
client: Stop caching media in client when doing setup If the media has been managed by a session media, it should not be cached in the client any longer. The GstRTSPSessionMedia object is now responsible for unpreparing the GstRTSPMedia object using gst_rtsp_media_unprepare(). Unprepare the media when finalizing the session media. https://bugzilla.gnome.org/show_bug.cgi?id=739112
2014-10-23 10:54:03 +00:00
/* don't cache media anymore */
client: refactor cleanup of cached media
2014-11-07 11:48:53 +00:00
clean_cached_media (client, FALSE);
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
}
ctx->sessmedia = sessmedia;
client: Configure transport after creating session media The default implementation of configure_client_transport() in rtsp-client uses the session media when it chooses channels for interleaved traffic. https://bugzilla.gnome.org/show_bug.cgi?id=739112
2014-10-24 10:04:54 +00:00
/* update the client transport */
if (!klass->configure_client_transport (client, ctx, ct))
goto unsupported_client_transport;
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* set in the session media transport */
stream: use the address managed by the stream Use the address managed by the stream for multicast. This allows us to have 1 multicast address for each stream. Because the address is now managed by the stream we don't have to pass it around anymore. Set the address pool on the streams.
2012-11-15 15:18:29 +00:00
trans = gst_rtsp_session_media_set_transport (sessmedia, stream, ct);
Initial import
2008-10-09 12:29:12 +00:00
rtsp-client: add stream transport to context We add the stream transport to the context so we can get the configured client stream transport in the setup request signal. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=738905
2014-10-20 22:40:59 +00:00
ctx->trans = trans;
client: store setup uri and use in PLAY response Store the uri used when doing the setup and use that in the PLAY response. fixes https://bugzilla.gnome.org/show_bug.cgi?id=715168
2013-11-28 16:47:18 +00:00
/* configure the url used to set this transport, this we will use when
* generating the response for the PLAY request */
gst_rtsp_stream_transport_set_url (trans, uri);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
/* configure keepalive for this transport */
rtsp: massive refactoring Make GObjects from the remaining simple structures. Remove GstRTSPSessionStream, it's not needed. Rename GstRTSPMediaStream -> GstRTSPStream: It is shorter Rename GstRTSPMediaTrans -> GstRTSPStreamTransport: It describes how a GstRTSPStream should be transported to a client. Rename GstRTSPMediaFactory::get_element -> create_element because that more accurately describes what it does. Make nice methods instead of poking in the structures. Move some methods inside the relevant object source code. Use GPtrArray to store objects instead of plain arrays, it is more natural and allows us to more easily clean up. Move the allocation of udp ports to the Stream object. The Stream object contains the elements needed to stream the media to a client. Improve the prepare and unprepare methods. Unprepare should now undo everything prepare did. Improve also async unprepare when doing EOS on shutdown. Make sure we always unprepare correctly.
2012-10-25 19:29:58 +00:00
gst_rtsp_stream_transport_set_keepalive (trans,
client: use right size for malloc
2010-03-16 17:33:23 +00:00
(GstRTSPKeepAliveFunc) do_keepalive, session, NULL);
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
if (ct->lower_transport == GST_RTSP_LOWER_TRANS_TCP) {
/* our callbacks to send data on this TCP connection */
gst_rtsp_stream_transport_set_callbacks (trans,
(GstRTSPSendFunc) do_send_data,
(GstRTSPSendFunc) do_send_data, client, NULL);
rtsp-server: Add support for buffer lists This adds new functions for passing buffer lists through the different layers without breaking API/ABI, and enables the appsink to actually provide buffer lists. This should already reduce CPU usage and potentially context switches a bit by passing a whole buffer list from the appsink instead of individual buffers. As a next step it would be necessary to a) Add support for a vector of data for the GstRTSPMessage body b) Add support for sending multiple messages at once to the GstRTSPWatch and let it be handled internally c) Adding API to GOutputStream that works like writev() Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-06-27 10:17:07 +00:00
gst_rtsp_stream_transport_set_list_callbacks (trans,
(GstRTSPSendListFunc) do_send_data_list,
(GstRTSPSendListFunc) do_send_data_list, client, NULL);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
stream: refactor TCP backpressure handling The previous implementation stopped sending TCP messages to all clients when a single one stopped consuming them, which obviously created problems for shared media. Instead, we now manage a backlog in stream-transport, and slow clients are removed once this backlog exceeds a maximum duration, currently hardcoded. Fixes #80
2019-10-15 17:08:32 +00:00
gst_rtsp_stream_transport_set_back_pressure_callback (trans,
(GstRTSPBackPressureFunc) do_check_back_pressure, client, NULL);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
g_hash_table_insert (priv->transports,
GINT_TO_POINTER (ct->interleaved.min), trans);
rtsp: Ref transports in hash table. Also ref streams for transports. This solves a crash when reciving a rtcp after teardown but before client finalize. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=740845
2014-11-28 11:45:14 +00:00
g_object_ref (trans);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
g_hash_table_insert (priv->transports,
GINT_TO_POINTER (ct->interleaved.max), trans);
rtsp: Ref transports in hash table. Also ref streams for transports. This solves a crash when reciving a rtcp after teardown but before client finalize. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=740845
2014-11-28 11:45:14 +00:00
g_object_ref (trans);
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
add_data_seq (client, ct->interleaved.min);
add_data_seq (client, ct->interleaved.max);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
}
rtsp: refactor configuration of transport Move the configuration of the transport to a place where it makes more sense.
2012-10-27 21:49:24 +00:00
/* create and serialize the server transport */
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
st = make_server_transport (client, media, ctx, ct);
Initial import
2008-10-09 12:29:12 +00:00
trans_str = gst_rtsp_transport_as_text (st);
Fix various leaks Fix some leaks.
2009-01-30 16:06:26 +00:00
gst_rtsp_transport_free (st);
Initial import
2008-10-09 12:29:12 +00:00
/* construct the response now */
code = GST_RTSP_STS_OK;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, code,
gst_rtsp_status_as_text (code), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_TRANSPORT,
client: use the response from the clientstate Create the response object only once and store in the client state. Make all methods use the state response,
2011-01-12 14:37:39 +00:00
trans_str);
Initial import
2008-10-09 12:29:12 +00:00
g_free (trans_str);
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (pipelined_request_id)
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_PIPELINED_REQUESTS,
pipelined_request_id);
if (ctx->request->type_data.request.version >= GST_RTSP_VERSION_2_0) {
GstClockTimeDiff seekable = gst_rtsp_media_seekable (media);
GString *media_properties = g_string_new (NULL);
if (seekable == -1)
g_string_append (media_properties,
"No-Seeking,Time-Progressing,Time-Duration=0.0");
else if (seekable == 0)
g_string_append (media_properties, "Beginning-Only");
else if (seekable == G_MAXINT64)
g_string_append (media_properties, "Random-Access");
else
g_string_append_printf (media_properties,
"Random-Access=%f, Unlimited, Immutable",
(gdouble) seekable / GST_SECOND);
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_MEDIA_PROPERTIES,
rtsp-client: fix leak adding headers gst_rtsp_message_add_header() makes a copy of the header, instead of taking ownership. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/204>
2021-05-05 00:45:19 +00:00
media_properties->str);
g_string_free (media_properties, TRUE);
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
/* TODO Check how Accept-Ranges should be filled */
gst_rtsp_message_add_header (ctx->request, GST_RTSP_HDR_ACCEPT_RANGES,
"npt, clock, smpte, clock");
}
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
Initial import
2008-10-09 12:29:12 +00:00
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
/* update the state */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
rtspstate = gst_rtsp_session_media_get_rtsp_state (sessmedia);
switch (rtspstate) {
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
case GST_RTSP_STATE_PLAYING:
case GST_RTSP_STATE_RECORDING:
case GST_RTSP_STATE_READY:
/* no state change */
break;
default:
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
gst_rtsp_session_media_set_rtsp_state (sessmedia, GST_RTSP_STATE_READY);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
break;
}
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_SETUP_REQUEST], 0, ctx);
gst_rtsp_media_unlock (media);
rtsp-client: Fix leaking of media in error cases With additional fixes by Kseniya Vasilchuk <vasilchukkseniia@gmail.com> and myself to make the media refcounting a bit easier to follow. https://bugzilla.gnome.org/show_bug.cgi?id=755632
2015-09-25 15:04:00 +00:00
g_object_unref (media);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
g_object_unref (session);
client: Add query to control path If the SETUP url contains a query it must be appended to the control path so that it matches any already created stream in the media. The query will also be appended to the session media path.
2013-10-01 12:04:17 +00:00
g_free (path);
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
g_free (control);
Rework the way we handle transports for streams Make the media accept an array of transports for the streams that we have configured for the play/pause requests. Implement server states for a client and its media. Require 0.10.22.1 (git HEAD) of gstreamer.
2009-02-03 18:32:38 +00:00
Initial import
2008-10-09 12:29:12 +00:00
return TRUE;
/* ERRORS */
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
no_uri:
{
GST_ERROR ("client %p: no uri", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
return FALSE;
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
no_transport:
Initial import
2008-10-09 12:29:12 +00:00
{
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
GST_ERROR ("client %p: no transport", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_UNSUPPORTED_TRANSPORT, ctx);
client: cleanup error paths
2014-04-03 10:57:13 +00:00
goto cleanup_path;
Initial import
2008-10-09 12:29:12 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
no_pool:
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
{
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
GST_ERROR ("client %p: no session pool configured", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
client: cleanup error paths
2014-04-03 10:57:13 +00:00
goto cleanup_path;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
}
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
media_not_found_no_reply:
{
GST_ERROR ("client %p: media '%s' not found", client, path);
/* error reply is already sent */
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
goto cleanup_session;
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
media_not_found:
{
GST_ERROR ("client %p: media '%s' not found", client, path);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
goto cleanup_session;
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
}
client: Do not read beyond end of path string If the setup was done without a control url, make sure we don't try to read the non-existing control string and crash.
2013-09-19 13:44:26 +00:00
control_not_found:
{
GST_ERROR ("client %p: no control in path '%s'", client, path);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
client: Do not read beyond end of path string If the setup was done without a control url, make sure we don't try to read the non-existing control string and crash.
2013-09-19 13:44:26 +00:00
g_object_unref (media);
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
goto cleanup_session;
client: Do not read beyond end of path string If the setup was done without a control url, make sure we don't try to read the non-existing control string and crash.
2013-09-19 13:44:26 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
stream_not_found:
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
{
rtsp-client: If we have a single-stream media and SETUP contains no control, use the one and only stream
2015-01-19 12:09:20 +00:00
GST_ERROR ("client %p: stream '%s' not found", client,
GST_STR_NULL (control));
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
g_object_unref (media);
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
goto cleanup_session;
rtsp: Handle the blocksize parameter Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679325
2012-07-03 11:26:30 +00:00
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_object_unref (media);
goto cleanup_path;
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
service_unavailable:
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
{
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
GST_ERROR ("client %p: can't create session", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
g_object_unref (media);
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
goto cleanup_session;
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
}
sessmedia_unavailable:
{
GST_ERROR ("client %p: can't create session media", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
rtsp-client: Fix leaking of media in error cases With additional fixes by Kseniya Vasilchuk <vasilchukkseniia@gmail.com> and myself to make the media refcounting a bit easier to follow. https://bugzilla.gnome.org/show_bug.cgi?id=755632
2015-09-25 15:04:00 +00:00
goto cleanup_transport;
rtsp: use AddressPool Remove the multicast_group property. Use the configured addresspool to allocate multicast addresses.
2012-11-14 16:23:59 +00:00
}
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
configure_media_failed_no_reply:
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
{
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
GST_ERROR ("client %p: configure_media failed", client);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
rtsp-client: Fix leaking of media in error cases With additional fixes by Kseniya Vasilchuk <vasilchukkseniia@gmail.com> and myself to make the media refcounting a bit easier to follow. https://bugzilla.gnome.org/show_bug.cgi?id=755632
2015-09-25 15:04:00 +00:00
g_object_unref (media);
client: add vmethod to configure media and streams Implement a vmethod that can be used to configure the media and the streams based on the current context. Handle the blocksize handling in the default handler. See https://bugzilla.gnome.org/show_bug.cgi?id=720667
2013-12-18 14:57:03 +00:00
/* error reply is already sent */
client: cleanup error paths
2014-04-03 10:57:13 +00:00
goto cleanup_session;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
}
Initial import
2008-10-09 12:29:12 +00:00
unsupported_transports:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: unsupported transports", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_UNSUPPORTED_TRANSPORT, ctx);
client: cleanup error paths
2014-04-03 10:57:13 +00:00
goto cleanup_transport;
Initial import
2008-10-09 12:29:12 +00:00
}
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
unsupported_client_transport:
Initial import
2008-10-09 12:29:12 +00:00
{
client: rework setup request a little Cache the media in DESCRIBE based on the longest matching path with the uri that we can find in the mount points. Rework the setup request a little to get the media from the session or from the longest matching path, this way we can derive the control string as everything after the path instead of hardcoding it. Find the stream based on the control string and only open a session when all this can be done.
2013-07-03 13:55:38 +00:00
GST_ERROR ("client %p: unsupported client transport", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_UNSUPPORTED_TRANSPORT, ctx);
client: cleanup error paths
2014-04-03 10:57:13 +00:00
goto cleanup_transport;
}
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
unsupported_mode:
{
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
GST_ERROR ("client %p: unsupported mode (media play: %d, media record: %d, "
"mode play: %d, mode record: %d)", client,
! !(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_PLAY),
! !(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_RECORD), ct->mode_play, ct->mode_record);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
send_generic_response (client, GST_RTSP_STS_UNSUPPORTED_TRANSPORT, ctx);
goto cleanup_transport;
}
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
unsupported_range_unit:
{
GST_ERROR ("Client %p: does not support any range format we support",
client);
send_generic_response (client, GST_RTSP_STS_NOT_IMPLEMENTED, ctx);
goto cleanup_transport;
}
client: parse the mikey response from the client Parse the mikey response from the client and update the policy for each SSRC.
2014-04-03 10:52:51 +00:00
keymgmt_error:
{
GST_ERROR ("client %p: keymgmt error", client);
send_generic_response (client, GST_RTSP_STS_KEY_MANAGEMENT_FAILURE, ctx);
goto cleanup_transport;
}
client: cleanup error paths
2014-04-03 10:57:13 +00:00
{
cleanup_transport:
client: don't leak transports
2012-06-14 07:59:06 +00:00
gst_rtsp_transport_free (ct);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
if (media) {
gst_rtsp_media_unlock (media);
rtsp-client: Fix leaking of media in error cases With additional fixes by Kseniya Vasilchuk <vasilchukkseniia@gmail.com> and myself to make the media refcounting a bit easier to follow. https://bugzilla.gnome.org/show_bug.cgi?id=755632
2015-09-25 15:04:00 +00:00
g_object_unref (media);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
}
client: cleanup error paths
2014-04-03 10:57:13 +00:00
cleanup_session:
client: manage media in session as a last step Once we manage a media in a session, we can't unmanage it anymore without destroying it. Therefore, first check everything before we manage the media, otherwise if something is wrong we have no way to unmanage the media. If we created a new session and something went wrong, remove the session again. Fixes a leak in the unit test.
2014-07-08 10:36:12 +00:00
if (new_session)
gst_rtsp_session_pool_remove (priv->session_pool, session);
rtsp-client: Fix leaking of session in error cases https://bugzilla.gnome.org/show_bug.cgi?id=755632
2016-08-02 12:08:22 +00:00
if (session)
g_object_unref (session);
client: cleanup error paths
2014-04-03 10:57:13 +00:00
cleanup_path:
client: Add query to control path If the SETUP url contains a query it must be appended to the control path so that it matches any already created stream in the media. The query will also be appended to the session media path.
2013-10-01 12:04:17 +00:00
g_free (path);
Make a mount point of "/" work correctly. As far as I can tell, this is neither explicitly allowed nor forbidden by RFC 7826. Meanwhile, URLs such as rtsp://<IP>:554 or rtsp://<IP>:554/ are in use in the wild (presumably with non-GStreamer servers). GStreamer's prior behavior was confusing, in that gst_rtsp_mount_points_add_factory() would appear to accept a mount path of "" or "/", but later connection attempts would fail with a "media not found" error. This commit makes a mount path of "/" work for either form of URL, while an empty mount path ("") is rejected and logs a warning. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/168>
2020-11-05 21:02:49 +00:00
g_free (control);
Initial import
2008-10-09 12:29:12 +00:00
return FALSE;
}
}
client: use right size for malloc
2010-03-16 17:33:23 +00:00
static GstSDPMessage *
create_sdp (GstRTSPClient * client, GstRTSPMedia * media)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
client: use right size for malloc
2010-03-16 17:33:23 +00:00
GstSDPMessage *sdp;
GstSDPInfo info;
const gchar *proto;
rtsp-client: Use a random session ID in the SDP RFC4566 Section 5.2 says that it should make the username, session id, nettype, addrtype and unicast address tuple globally unique. Always using 1188340656180883 is not going to guarantee that: https://xkcd.com/221/ Instead let's create a 64 bit random number, which at least brings us closer to the goal of global uniqueness. https://tools.ietf.org/html/rfc4566#section-5.2
2015-01-18 18:08:36 +00:00
guint64 session_id_tmp;
gchar session_id[21];
client: use right size for malloc
2010-03-16 17:33:23 +00:00
gst_sdp_message_new (&sdp);
/* some standard things first */
gst_sdp_message_set_version (sdp, "0");
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (priv->is_ipv6)
client: use right size for malloc
2010-03-16 17:33:23 +00:00
proto = "IP6";
else
proto = "IP4";
rtsp-client: Use a random session ID in the SDP RFC4566 Section 5.2 says that it should make the username, session id, nettype, addrtype and unicast address tuple globally unique. Always using 1188340656180883 is not going to guarantee that: https://xkcd.com/221/ Instead let's create a 64 bit random number, which at least brings us closer to the goal of global uniqueness. https://tools.ietf.org/html/rfc4566#section-5.2
2015-01-18 18:08:36 +00:00
session_id_tmp = (((guint64) g_random_int ()) << 32) | g_random_int ();
g_snprintf (session_id, sizeof (session_id), "%" G_GUINT64_FORMAT,
session_id_tmp);
gst_sdp_message_set_origin (sdp, "-", session_id, "1", "IN", proto,
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->server_ip);
client: use right size for malloc
2010-03-16 17:33:23 +00:00
gst_sdp_message_set_session_name (sdp, "Session streamed with GStreamer");
gst_sdp_message_set_information (sdp, "rtsp-server");
gst_sdp_message_add_time (sdp, "0", "0", NULL);
gst_sdp_message_add_attribute (sdp, "tool", "GStreamer");
gst_sdp_message_add_attribute (sdp, "type", "broadcast");
gst_sdp_message_add_attribute (sdp, "control", "*");
sdp: cleanup sdp info We don't need to pass the proto, we can more easily check a boolean. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702063
2013-06-14 14:05:59 +00:00
info.is_ipv6 = priv->is_ipv6;
info.server_ip = priv->server_ip;
client: use right size for malloc
2010-03-16 17:33:23 +00:00
/* create an SDP for the media object */
media: add setup_sdp vmethod gst/rtsp-server/rtsp-media.[ch]: added setup_sdp vmethod and public gst_rtsp_media_setup_sdp. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=720155
2013-12-18 15:37:27 +00:00
if (!gst_rtsp_media_setup_sdp (media, sdp, &info))
client: use right size for malloc
2010-03-16 17:33:23 +00:00
goto no_sdp;
return sdp;
/* ERRORS */
no_sdp:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: could not create SDP", client);
client: use right size for malloc
2010-03-16 17:33:23 +00:00
gst_sdp_message_free (sdp);
return NULL;
}
}
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
/* for the describe we must generate an SDP */
Initial import
2008-10-09 12:29:12 +00:00
static gboolean
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_describe_request (GstRTSPClient * client, GstRTSPContext * ctx)
Initial import
2008-10-09 12:29:12 +00:00
{
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
GstRTSPClientPrivate *priv = client->priv;
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
GstRTSPResult res;
Initial import
2008-10-09 12:29:12 +00:00
GstSDPMessage *sdp;
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
guint i;
gchar *path, *str;
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
GstRTSPMedia *media;
rtsp-client: make create_sdp virtual method Fixes https://bugzilla.gnome.org/show_bug.cgi?id=680173
2012-07-18 13:54:49 +00:00
GstRTSPClientClass *klass;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
rtsp-client: make create_sdp virtual method Fixes https://bugzilla.gnome.org/show_bug.cgi?id=680173
2012-07-18 13:54:49 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!ctx->uri)
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
goto no_uri;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_DESCRIBE_REQUEST],
0, ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
/* check what kind of format is accepted, we don't really do anything with it
* and always return SDP for now. */
rtsp-client: Fix iteration Wouldn't even enter the code block otherwise (i++ was used as the check and not the postfix).
2013-12-19 13:24:54 +00:00
for (i = 0;; i++) {
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
gchar *accept;
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
res =
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_ACCEPT,
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
&accept, i);
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
if (res == GST_RTSP_ENOTIMPL)
break;
if (g_ascii_strcasecmp (accept, "application/sdp") == 0)
break;
}
Initial import
2008-10-09 12:29:12 +00:00
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
if (!priv->mount_points)
goto no_mount_points;
if (!(path = gst_rtsp_mount_points_make_path (priv->mount_points, ctx->uri)))
goto no_path;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
/* find the media object for the uri */
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
if (!(media = find_media (client, ctx, path, NULL)))
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
goto no_media;
Cleanups and doc updates Add some more documentation and do some minor cleanups here and there.
2009-01-22 17:35:17 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_lock (media);
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if (!(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_PLAY))
goto unsupported_mode;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
client: use right size for malloc
2010-03-16 17:33:23 +00:00
/* create an SDP for the media object on this client */
rtsp-client: make create_sdp virtual method Fixes https://bugzilla.gnome.org/show_bug.cgi?id=680173
2012-07-18 13:54:49 +00:00
if (!(sdp = klass->create_sdp (client, media)))
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
goto no_sdp;
Initial import
2008-10-09 12:29:12 +00:00
client: suspend after SDP and unsuspend before PLAYING Based on patches by Ognyan Tonchev <ognyan@axis.com> Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711257
2013-11-28 13:14:35 +00:00
/* we suspend after the describe */
gst_rtsp_media_suspend (media);
Cleanup of sessions and more Fix the refcounting of media and sessions in the client. Properly clean up the session data when the client performs a teardown. Add Server header to responses. Allow for multiple uri setups in one session. Add Range header to the PLAY response and add the range attribute to the SDP message. Fix the session pool remove method, it used the wrong key in the hashtable. Also give the ownership of the sessionid to the session object.
2009-02-04 16:00:42 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, GST_RTSP_STS_OK,
gst_rtsp_status_as_text (GST_RTSP_STS_OK), ctx->request);
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_CONTENT_TYPE,
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
"application/sdp");
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
Add a little comment Add some comment about the content-base header.
2009-01-29 16:19:21 +00:00
/* content base for some clients that might screw up creating the setup uri */
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
str = make_base_url (client, ctx->uri, path);
g_free (path);
client: use right size for malloc
2010-03-16 17:33:23 +00:00
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
GST_INFO ("adding content-base: %s", str);
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_CONTENT_BASE, str);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
Initial import
2008-10-09 12:29:12 +00:00
/* add SDP to the response body */
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
str = gst_sdp_message_as_text (sdp);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_take_body (ctx->response, (guint8 *) str, strlen (str));
Initial import
2008-10-09 12:29:12 +00:00
gst_sdp_message_free (sdp);
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
Initial import
2008-10-09 12:29:12 +00:00
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_DESCRIBE_REQUEST],
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
Initial import
2008-10-09 12:29:12 +00:00
return TRUE;
/* ERRORS */
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
return FALSE;
}
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
no_uri:
{
GST_ERROR ("client %p: no uri", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
return FALSE;
}
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
no_mount_points:
{
GST_ERROR ("client %p: no mount points configured", client);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
return FALSE;
}
no_path:
{
GST_ERROR ("client %p: can't find path for url", client);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
return FALSE;
}
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
no_media:
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no media", client);
client: map url to path only in describe Only map the request url to a path in the DESCRIBE method. The SDP then contains the base and control urls that should be used to SETUP/PAUSE/ PLAY/TEARDOWN the media.
2013-09-17 10:21:02 +00:00
g_free (path);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
/* error reply is already sent */
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
return FALSE;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
unsupported_mode:
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
{
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
GST_ERROR ("client %p: media does not support DESCRIBE", client);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_ALLOWED, ctx);
g_free (path);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
g_object_unref (media);
return FALSE;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
no_sdp:
Handle state change failures better Handle state change failures better when changing the state of the pipeline to determine the SDP.
2009-01-22 15:53:16 +00:00
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: can't create SDP", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
client: Fix RTPInfo header Refactor the method to make the content_base. Use the content-base and the control url to construct the RTPInfo url.
2013-09-17 12:39:44 +00:00
g_free (path);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
g_object_unref (media);
Handle state change failures better Handle state change failures better when changing the state of the pipeline to determine the SDP.
2009-01-22 15:53:16 +00:00
return FALSE;
}
Initial import
2008-10-09 12:29:12 +00:00
}
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
static gboolean
handle_sdp (GstRTSPClient * client, GstRTSPContext * ctx, GstRTSPMedia * media,
GstSDPMessage * sdp)
{
GstRTSPClientPrivate *priv = client->priv;
GstRTSPThread *thread;
/* create an SDP for the media object */
if (!gst_rtsp_media_handle_sdp (media, sdp))
goto unhandled_sdp;
thread = gst_rtsp_thread_pool_get_thread (priv->thread_pool,
GST_RTSP_THREAD_TYPE_MEDIA, ctx);
if (thread == NULL)
goto no_thread;
/* prepare the media */
if (!gst_rtsp_media_prepare (media, thread))
goto no_prepare;
return TRUE;
/* ERRORS */
unhandled_sdp:
{
GST_ERROR ("client %p: could not handle SDP", client);
return FALSE;
}
no_thread:
{
GST_ERROR ("client %p: can't create thread", client);
return FALSE;
}
no_prepare:
{
GST_ERROR ("client %p: can't prepare media", client);
return FALSE;
}
}
static gboolean
handle_announce_request (GstRTSPClient * client, GstRTSPContext * ctx)
{
GstRTSPClientPrivate *priv = client->priv;
GstRTSPClientClass *klass;
GstSDPResult sres;
GstSDPMessage *sdp;
GstRTSPMedia *media;
gchar *path, *cont = NULL;
guint8 *data;
guint size;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
guint i, n_streams;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
klass = GST_RTSP_CLIENT_GET_CLASS (client);
if (!ctx->uri)
goto no_uri;
if (!priv->mount_points)
goto no_mount_points;
/* check if reply is SDP */
gst_rtsp_message_get_header (ctx->request, GST_RTSP_HDR_CONTENT_TYPE, &cont,
0);
/* could not be set but since the request returned OK, we assume it
* was SDP, else check it. */
if (cont) {
rtsp-client: fix awkward if clause
2015-02-08 12:08:36 +00:00
if (g_ascii_strcasecmp (cont, "application/sdp") != 0)
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
goto wrong_content_type;
}
/* get message body and parse as SDP */
gst_rtsp_message_get_body (ctx->request, &data, &size);
if (data == NULL || size == 0)
goto no_message;
GST_DEBUG ("client %p: parse SDP...", client);
gst_sdp_message_new (&sdp);
sres = gst_sdp_message_parse_buffer (data, size, sdp);
if (sres != GST_SDP_OK)
goto sdp_parse_failed;
if (!(path = gst_rtsp_mount_points_make_path (priv->mount_points, ctx->uri)))
goto no_path;
/* find the media object for the uri */
if (!(media = find_media (client, ctx, path, NULL)))
goto no_media;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
ctx->media = media;
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_lock (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_ANNOUNCE_REQUEST],
0, ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if (!(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_RECORD))
goto unsupported_mode;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
/* Tell client subclass about the media */
if (!klass->handle_sdp (client, ctx, media, sdp))
goto unhandled_sdp;
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
gst_rtsp_message_init_response (ctx->response, GST_RTSP_STS_OK,
gst_rtsp_status_as_text (GST_RTSP_STS_OK), ctx->request);
n_streams = gst_rtsp_media_n_streams (media);
for (i = 0; i < n_streams; i++) {
GstRTSPStream *stream = gst_rtsp_media_get_stream (media, i);
rtsp-client: Generate correct URI for MIKEY in ANNOUNCE responses Instead of hardcoding the URI, take the actual URI (and especially the correct port) from the RTSP context. Fixes #84
2019-10-17 10:15:42 +00:00
gchar *uri, *location, *keymgmt;
uri = gst_rtsp_url_get_request_uri (ctx->uri);
location = g_strdup_printf ("%s/stream=%d", uri, i);
keymgmt = stream_make_keymgmt (client, location, stream);
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
if (keymgmt)
gst: Run everything through gst-indent again
2018-04-04 07:06:06 +00:00
gst_rtsp_message_take_header (ctx->response, GST_RTSP_HDR_KEYMGMT,
keymgmt);
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
g_free (location);
rtsp-client: Generate correct URI for MIKEY in ANNOUNCE responses Instead of hardcoding the URI, take the actual URI (and especially the correct port) from the RTSP context. Fixes #84
2019-10-17 10:15:42 +00:00
g_free (uri);
rtsp-client: Send KeyMgmt header in ANNOUNCE response When sending back an encrypted RTCP back channel, it is useful for the client to know the encryption key. https://bugzilla.gnome.org/show_bug.cgi?id=794813
2018-03-29 14:12:26 +00:00
}
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
/* we suspend after the announce */
gst_rtsp_media_suspend (media);
send_message (client, ctx, ctx->response, FALSE);
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_ANNOUNCE_REQUEST],
0, ctx);
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
g_free (path);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return TRUE;
no_uri:
{
GST_ERROR ("client %p: no uri", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
return FALSE;
}
no_mount_points:
{
GST_ERROR ("client %p: no mount points configured", client);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
return FALSE;
}
no_path:
{
GST_ERROR ("client %p: can't find path for url", client);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
wrong_content_type:
{
GST_ERROR ("client %p: unknown content type", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
return FALSE;
}
no_message:
{
GST_ERROR ("client %p: can't find SDP message", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
return FALSE;
}
sdp_parse_failed:
{
GST_ERROR ("client %p: failed to parse SDP message", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
no_media:
{
GST_ERROR ("client %p: no media", client);
g_free (path);
/* error reply is already sent */
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
gst_sdp_message_free (sdp);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
g_object_unref (media);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
return FALSE;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
unsupported_mode:
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
{
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
GST_ERROR ("client %p: media does not support ANNOUNCE", client);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_ALLOWED, ctx);
g_free (path);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
g_object_unref (media);
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
unhandled_sdp:
{
GST_ERROR ("client %p: can't handle SDP", client);
send_generic_response (client, GST_RTSP_STS_UNSUPPORTED_MEDIA_TYPE, ctx);
g_free (path);
rtsp-client: Lock shared media For shared media we got race conditions. Concurrently rtsp clients might suspend or unsuspend the shared media and thus change the state without the clients expecting that. By introducing a lock that can be taken by callers such as rtsp_client one can force rtsp clients calling, eg. PLAY, SETUP and that uses shared media, to handle the media sequentially thus allowing one client to finish its rtsp call before another client calls on the same media. https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/86 Fixes #86
2019-10-16 13:20:54 +00:00
gst_rtsp_media_unlock (media);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
g_object_unref (media);
rtsp-client: fix a couple of leaks in handle_announce
2015-01-21 16:32:44 +00:00
gst_sdp_message_free (sdp);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
return FALSE;
}
}
static gboolean
handle_record_request (GstRTSPClient * client, GstRTSPContext * ctx)
{
GstRTSPSession *session;
GstRTSPClientClass *klass;
GstRTSPSessionMedia *sessmedia;
GstRTSPMedia *media;
GstRTSPUrl *uri;
GstRTSPState rtspstate;
gchar *path;
gint matched;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
GPtrArray *transports;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
if (!(session = ctx->session))
goto no_session;
if (!(uri = ctx->uri))
goto no_uri;
klass = GST_RTSP_CLIENT_GET_CLASS (client);
path = klass->make_path_from_uri (client, uri);
/* get a handle to the configuration of the media in the session */
sessmedia = gst_rtsp_session_get_media (session, path, &matched);
if (!sessmedia)
goto not_found;
if (path[matched] != '\0')
goto no_aggregate;
g_free (path);
ctx->sessmedia = sessmedia;
ctx->media = media = gst_rtsp_session_media_get_media (sessmedia);
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_RECORD_REQUEST], 0,
ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
if (!(gst_rtsp_media_get_transport_mode (media) &
GST_RTSP_TRANSPORT_MODE_RECORD))
goto unsupported_mode;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
/* the session state must be playing or ready */
rtspstate = gst_rtsp_session_media_get_rtsp_state (sessmedia);
if (rtspstate != GST_RTSP_STATE_PLAYING && rtspstate != GST_RTSP_STATE_READY)
goto invalid_state;
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
/* update the pipeline */
transports = gst_rtsp_session_media_get_transports (sessmedia);
if (!gst_rtsp_media_complete_pipeline (media, transports)) {
g_ptr_array_unref (transports);
goto pipeline_error;
}
g_ptr_array_unref (transports);
/* in record we first unsuspend, media could be suspended from SDP or PAUSED */
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
if (!gst_rtsp_media_unsuspend (media))
goto unsuspend_failed;
gst_rtsp_message_init_response (ctx->response, GST_RTSP_STS_OK,
gst_rtsp_status_as_text (GST_RTSP_STS_OK), ctx->request);
send_message (client, ctx, ctx->response, FALSE);
/* start playing after sending the response */
gst_rtsp_session_media_set_state (sessmedia, GST_STATE_PLAYING);
gst_rtsp_session_media_set_rtsp_state (sessmedia, GST_RTSP_STATE_PLAYING);
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_RECORD_REQUEST], 0,
ctx);
return TRUE;
/* ERRORS */
no_session:
{
GST_ERROR ("client %p: no session", client);
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
return FALSE;
}
no_uri:
{
GST_ERROR ("client %p: no uri supplied", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
return FALSE;
}
not_found:
{
GST_ERROR ("client %p: media not found", client);
send_generic_response (client, GST_RTSP_STS_NOT_FOUND, ctx);
return FALSE;
}
no_aggregate:
{
GST_ERROR ("client %p: no aggregate path %s", client, path);
send_generic_response (client,
GST_RTSP_STS_ONLY_AGGREGATE_OPERATION_ALLOWED, ctx);
g_free (path);
return FALSE;
}
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
return FALSE;
}
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
unsupported_mode:
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
{
rtsp-media: Use flags to distinguish between PLAY and RECORD media
2015-01-28 17:54:01 +00:00
GST_ERROR ("client %p: media does not support RECORD", client);
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_ALLOWED, ctx);
return FALSE;
}
invalid_state:
{
GST_ERROR ("client %p: not PLAYING or READY", client);
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_VALID_IN_THIS_STATE,
Dynamically reconfigure pipeline in PLAY based on transports The initial pipeline does not contain specific transport elements. The receiver and the sender parts are added after PLAY. If the media is shared, the streams are dynamically reconfigured after each PLAY. https://bugzilla.gnome.org/show_bug.cgi?id=788340
2017-10-17 08:44:33 +00:00
ctx);
return FALSE;
}
pipeline_error:
{
GST_ERROR ("client %p: failed to configure the pipeline", client);
send_generic_response (client, GST_RTSP_STS_METHOD_NOT_VALID_IN_THIS_STATE,
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
ctx);
return FALSE;
}
unsuspend_failed:
{
GST_ERROR ("client %p: unsuspend failed", client);
send_generic_response (client, GST_RTSP_STS_SERVICE_UNAVAILABLE, ctx);
return FALSE;
}
}
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
static gboolean
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
handle_options_request (GstRTSPClient * client, GstRTSPContext * ctx,
GstRTSPVersion version)
Initial import
2008-10-09 12:29:12 +00:00
{
GstRTSPMethod options;
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
gchar *str;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
GstRTSPStatusCode sig_result;
Initial import
2008-10-09 12:29:12 +00:00
options = GST_RTSP_DESCRIBE |
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
GST_RTSP_OPTIONS |
GST_RTSP_PAUSE |
GST_RTSP_PLAY |
GST_RTSP_SETUP |
GST_RTSP_GET_PARAMETER | GST_RTSP_SET_PARAMETER | GST_RTSP_TEARDOWN;
Initial import
2008-10-09 12:29:12 +00:00
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (version < GST_RTSP_VERSION_2_0) {
options |= GST_RTSP_RECORD;
options |= GST_RTSP_ANNOUNCE;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
str = gst_rtsp_options_as_text (options);
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_init_response (ctx->response, GST_RTSP_STS_OK,
gst_rtsp_status_as_text (GST_RTSP_STS_OK), ctx->request);
Cleanups and doc updates Add some more documentation and do some minor cleanups here and there.
2009-01-22 17:35:17 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
gst_rtsp_message_add_header (ctx->response, GST_RTSP_HDR_PUBLIC, str);
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
g_free (str);
Initial import
2008-10-09 12:29:12 +00:00
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_PRE_OPTIONS_REQUEST], 0,
ctx, &sig_result);
if (sig_result != GST_RTSP_STS_OK) {
goto sig_failed;
}
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
send_message (client, ctx, ctx->response, FALSE);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_OPTIONS_REQUEST],
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
0, ctx);
rtsp-client: add signals for rtsp requests (fixes #683287)
2012-09-03 17:48:14 +00:00
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
return TRUE;
client: emit signal in the beginning of each rtsp request These signals let the application validate the requests, configure the media/stream in a certain way and also generate error status code in case of error or bad request. https://bugzilla.gnome.org/show_bug.cgi?id=758062
2016-08-24 09:39:13 +00:00
/* ERRORS */
sig_failed:
{
GST_ERROR ("client %p: pre signal returned error: %s", client,
gst_rtsp_status_as_text (sig_result));
send_generic_response (client, sig_result, ctx);
gst_rtsp_message_free (ctx->response);
return FALSE;
}
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
}
Initial import
2008-10-09 12:29:12 +00:00
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
/* remove duplicate and trailing '/' */
static void
client: fix typo
2010-12-28 11:18:41 +00:00
sanitize_uri (GstRTSPUrl * uri)
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
{
gint i, len;
gchar *s, *d;
gboolean have_slash, prev_slash;
s = d = uri->abspath;
len = strlen (uri->abspath);
prev_slash = FALSE;
for (i = 0; i < len; i++) {
have_slash = s[i] == '/';
*d = s[i];
if (!have_slash || !prev_slash)
d++;
prev_slash = have_slash;
}
len = d - uri->abspath;
/* don't remove the first slash if that's the only thing left */
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
if (len > 1 && *(d - 1) == '/')
Reorganize things, prepare for media sharing Added various other test server examples Move the SDP message generation to a separate helper. Refactor common code for finding the session. Add content-base for realplayer compatibility Clean up request uris before processing for better vlc compatibility. Move prerolling and pipeline construction to the RTSPMedia object. Use multiudpsink for future pipeline reuse.
2009-01-29 12:31:27 +00:00
d--;
*d = '\0';
Initial import
2008-10-09 12:29:12 +00:00
}
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
/* is called when the session is removed from its session pool. */
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
static void
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
client_session_removed (GstRTSPSessionPool * pool, GstRTSPSession * session,
GstRTSPClient * client)
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
{
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
GSource *timer_src;
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
GST_INFO ("client %p: session %p removed", client, session);
client: protect sessions with lock Protect the list of sessions with the lock. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 12:41:14 +00:00
g_mutex_lock (&priv->lock);
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
client_unwatch_session (client, session, NULL);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
if (!priv->sessions && priv->rtsp_ctrl_timeout == NULL) {
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
if (priv->post_session_timeout > 0) {
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
GWeakRef *client_weak_ref = g_new (GWeakRef, 1);
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
timer_src = g_timeout_source_new_seconds (RTSP_CTRL_CB_INTERVAL);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
g_weak_ref_init (client_weak_ref, client);
g_source_set_callback (timer_src, rtsp_ctrl_timeout_cb, client_weak_ref,
rtsp_ctrl_timeout_destroy_notify);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
priv->rtsp_ctrl_timeout_cnt = 0;
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
g_source_attach (timer_src, priv->watch_context);
priv->rtsp_ctrl_timeout = timer_src;
GST_DEBUG ("rtsp control setting up connection timeout %p.",
priv->rtsp_ctrl_timeout);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
g_mutex_unlock (&priv->lock);
} else if (priv->post_session_timeout == 0) {
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
g_mutex_unlock (&priv->lock);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
gst_rtsp_client_close (client);
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
} else {
g_mutex_unlock (&priv->lock);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
}
rtsp-client: Combine the pre-session and post-session timeout They previously used the same state but different mechanisms and functions, which was difficult to follow, error prone and simply confusing. Also adjust the test for the post-session timeout a bit to be less racy now that the timing has slightly changed. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-27 20:25:22 +00:00
} else {
g_mutex_unlock (&priv->lock);
rtsp-client: add property post-session-timeout This is a TCP connection timeout for client connections, in seconds. If a positive value is set for this property, the client connection will be kept alive for this amount of seconds after the last session timeout. For negative values of this property the connection timeout handling is delegated to the system (just as it was before). Fixes #83
2019-12-17 15:08:19 +00:00
}
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
}
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
/* Check for Require headers. Returns TRUE if there are no Require headers,
* otherwise lets the application decide which headers are supported.
* By default all headers are unsupported.
* If there are unsupported options, FALSE will be returned together with
* a newly-allocated string of (comma-separated) unsupported options in
* the unsupported_reqs variable.
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
*
* There may be multiple Require headers, but we must send one single
* Unsupported header with all the unsupported options as response. If
* an incoming Require header contained a comma-separated list of options
* GstRtspConnection will already have split that list up into multiple
* headers.
*/
static gboolean
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
check_request_requirements (GstRTSPContext * ctx, gchar ** unsupported_reqs)
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
{
GstRTSPResult res;
GPtrArray *arr = NULL;
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
GstRTSPMessage *msg = ctx->request;
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
gchar *reqs = NULL;
gint i;
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
gchar *sig_result = NULL;
gboolean result = TRUE;
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
i = 0;
do {
res = gst_rtsp_message_get_header (msg, GST_RTSP_HDR_REQUIRE, &reqs, i++);
if (res == GST_RTSP_ENOTIMPL)
break;
if (arr == NULL)
arr = g_ptr_array_new_with_free_func ((GDestroyNotify) g_free);
g_ptr_array_add (arr, g_strdup (reqs));
}
while (TRUE);
/* if we don't have any Require headers at all, all is fine */
if (i == 1)
return TRUE;
/* otherwise we've now processed at all the Require headers */
g_ptr_array_add (arr, NULL);
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
g_signal_emit (ctx->client,
gst_rtsp_client_signals[SIGNAL_CHECK_REQUIREMENTS], 0, ctx,
(gchar **) arr->pdata, &sig_result);
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
if (sig_result == NULL) {
/* no supported options, just report all of the required ones as
* unsupported */
*unsupported_reqs = g_strjoinv (", ", (gchar **) arr->pdata);
result = FALSE;
goto done;
}
if (strlen (sig_result) == 0)
g_free (sig_result);
else {
*unsupported_reqs = sig_result;
result = FALSE;
}
done:
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
g_ptr_array_unref (arr);
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
return result;
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
}
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
static void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
handle_request (GstRTSPClient * client, GstRTSPMessage * request)
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPMethod method;
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
const gchar *uristr;
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
GstRTSPUrl *uri = NULL;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPVersion version;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
GstRTSPResult res;
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
GstRTSPSession *session = NULL;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
GstRTSPContext sctx = { NULL }, *ctx;
client: use the response from the clientstate Create the response object only once and store in the client state. Make all methods use the state response,
2011-01-12 14:37:39 +00:00
GstRTSPMessage response = { 0 };
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
gchar *unsupported_reqs = NULL;
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
gchar *sessid = NULL, *pipelined_request_id = NULL;
Initial import
2008-10-09 12:29:12 +00:00
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (!(ctx = gst_rtsp_context_get_current ())) {
ctx = &sctx;
ctx->auth = priv->auth;
gst_rtsp_context_push_current (ctx);
client: support pushed context in handle_request If we already have a pushed state, reuse it and add our own things. This makes it easier to write tests.
2013-07-15 09:57:49 +00:00
}
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->conn = priv->connection;
ctx->client = client;
ctx->request = request;
ctx->response = &response;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
client: dump rtsp message only if debug threshold is higher than GST_LEVEL_LOG
2009-11-21 00:17:25 +00:00
if (gst_debug_category_get_threshold (rtsp_client_debug) >= GST_LEVEL_LOG) {
gst_rtsp_message_dump (request);
}
Initial import
2008-10-09 12:29:12 +00:00
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
gst_rtsp_message_parse_request (request, &method, &uristr, &version);
More improvements Rename GstRTSPMediaBin to GstRTSPMedia Parse the request url into a GstRTSPUri object and pass this object to the various handlers and methods that require the uri.
2009-01-22 16:58:19 +00:00
media: also handle prepare and range in suspended state When we are suspended, we are already prepared. We can get the range in the suspended state.
2013-11-29 09:53:08 +00:00
GST_INFO ("client %p: received a request %s %s %s", client,
gst_rtsp_method_as_text (method), uristr,
gst_rtsp_version_as_text (version));
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
/* we can only handle 1.0 requests */
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (version != GST_RTSP_VERSION_1_0 && version != GST_RTSP_VERSION_2_0)
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto not_supported;
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->method = method;
Initial import
2008-10-09 12:29:12 +00:00
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
/* we always try to parse the url first */
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
if (strcmp (uristr, "*") == 0) {
/* special case where we have * as uri, keep uri = NULL */
client: allow absolute path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711689
2013-11-12 09:55:14 +00:00
} else if (gst_rtsp_url_parse (uristr, &uri) != GST_RTSP_OK) {
/* check if the uristr is an absolute path <=> scheme and host information
* is missing */
gchar *scheme;
scheme = g_uri_parse_scheme (uristr);
if (scheme == NULL && g_str_has_prefix (uristr, "/")) {
gchar *absolute_uristr = NULL;
GST_WARNING_OBJECT (client, "request doesn't contain absolute url");
if (priv->server_ip == NULL) {
GST_WARNING_OBJECT (client, "host information missing");
goto bad_request;
}
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
absolute_uristr =
g_strdup_printf ("rtsp://%s%s", priv->server_ip, uristr);
client: allow absolute path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=711689
2013-11-12 09:55:14 +00:00
GST_DEBUG_OBJECT (client, "absolute url: %s", absolute_uristr);
if (gst_rtsp_url_parse (absolute_uristr, &uri) != GST_RTSP_OK) {
g_free (absolute_uristr);
goto bad_request;
}
g_free (absolute_uristr);
} else {
g_free (scheme);
goto bad_request;
}
}
Initial import
2008-10-09 12:29:12 +00:00
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
/* get the session if there is any */
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
res = gst_rtsp_message_get_header (request, GST_RTSP_HDR_PIPELINED_REQUESTS,
&pipelined_request_id, 0);
if (res == GST_RTSP_OK) {
sessid = g_hash_table_lookup (client->priv->pipelined_requests,
pipelined_request_id);
if (!sessid)
res = GST_RTSP_ERROR;
}
if (res != GST_RTSP_OK)
res =
gst_rtsp_message_get_header (request, GST_RTSP_HDR_SESSION, &sessid, 0);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
if (res == GST_RTSP_OK) {
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (priv->session_pool == NULL)
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
goto no_pool;
Initial import
2008-10-09 12:29:12 +00:00
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
/* we had a session in the request, find it again */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if (!(session = gst_rtsp_session_pool_find (priv->session_pool, sessid)))
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
goto session_not_found;
Initial import
2008-10-09 12:29:12 +00:00
Remove weak ref on the session in teardown We need to remove our weakref from the session when we do a teardown because else we close the TCP connection prematurely.
2009-03-11 19:03:06 +00:00
/* we add the session to the client list of watched sessions. When a session
* disappears because it times out, we will be notified. If all sessions are
* gone, we will close the connection */
Do some more session cleanup Make session timeout kill the TCP connection that currently watches the session. Remove the client timeout property.
2009-03-11 18:38:06 +00:00
client_watch_session (client, session);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
}
Move the connection code in one place Add some comments
2009-01-20 12:57:47 +00:00
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
/* sanitize the uri */
client: handle asterisk as path in requests Fixes https://bugzilla.gnome.org/show_bug.cgi?id=701266
2013-06-20 09:17:29 +00:00
if (uri)
sanitize_uri (uri);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx->uri = uri;
ctx->session = session;
rtsp-server: Pass ClientState structure arround Pass the collected information for the ongoing request in a GstRTSPClientState structure that we can then pass around to simplify the method arguments. This will also be handy when we implement logging functionality.
2011-01-12 12:16:08 +00:00
auth: simplify auth checks Remove client from methods, it's now in the state Perform the check specified by the string, use the information from the thread local context.
2013-07-09 14:04:35 +00:00
if (!gst_rtsp_auth_check (GST_RTSP_AUTH_CHECK_URL))
goto not_authorized;
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
/* handle any 'Require' headers */
rtsp-client: allow application to decide what requirements are supported Add "check-requirements" signal and vfunc to allow application (and subclasses) to check the requirements. Based on patch from Hyunjun Ko <zzoon.ko@samsung.com> https://bugzilla.gnome.org/show_bug.cgi?id=749417
2015-06-18 11:12:04 +00:00
if (!check_request_requirements (ctx, &unsupported_reqs))
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
goto unsupported_requirement;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
/* now see what is asked and dispatch to a dedicated handler */
switch (method) {
case GST_RTSP_OPTIONS:
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
priv->version = version;
handle_options_request (client, ctx, version);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
case GST_RTSP_DESCRIBE:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_describe_request (client, ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
case GST_RTSP_SETUP:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_setup_request (client, ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
case GST_RTSP_PLAY:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_play_request (client, ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
case GST_RTSP_PAUSE:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_pause_request (client, ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
case GST_RTSP_TEARDOWN:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_teardown_request (client, ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
break;
client: replay OK to GET/SET_PARAMETER Some clients (vlc) use GET/SET_PARAMETER to keep the TCP session open. Make it so that we return OK for those requests.
2009-05-26 15:25:59 +00:00
case GST_RTSP_SET_PARAMETER:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_set_param_request (client, ctx);
rtsp: prepare for handling GET/SET_PARAMETER Add helper functions to handle GET/SET_PARAMETER. Reply with an error when there is a body now. Fix return codes of handlers.
2009-06-09 22:01:07 +00:00
break;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
case GST_RTSP_GET_PARAMETER:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
handle_get_param_request (client, ctx);
client: replay OK to GET/SET_PARAMETER Some clients (vlc) use GET/SET_PARAMETER to keep the TCP session open. Make it so that we return OK for those requests.
2009-05-26 15:25:59 +00:00
break;
case GST_RTSP_ANNOUNCE:
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (version >= GST_RTSP_VERSION_2_0)
goto invalid_command_for_version;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
handle_announce_request (client, ctx);
break;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
case GST_RTSP_RECORD:
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
if (version >= GST_RTSP_VERSION_2_0)
goto invalid_command_for_version;
Add initial support for RECORD We currently only support media that is RECORD or PLAY only, not both at once. https://bugzilla.gnome.org/show_bug.cgi?id=743175
2015-01-09 11:40:47 +00:00
handle_record_request (client, ctx);
break;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
case GST_RTSP_REDIRECT:
client: improve debug
2012-11-27 11:33:02 +00:00
goto not_implemented;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
case GST_RTSP_INVALID:
default:
client: improve debug
2012-11-27 11:33:02 +00:00
goto bad_request;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
}
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
done:
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
if (ctx == &sctx)
gst_rtsp_context_pop_current (ctx);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
if (session)
g_object_unref (session);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
if (uri)
gst_rtsp_url_free (uri);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
return;
Initial import
2008-10-09 12:29:12 +00:00
/* ERRORS */
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
not_supported:
{
GST_ERROR ("client %p: version %d not supported", client, version);
send_generic_response (client, GST_RTSP_STS_RTSP_VERSION_NOT_SUPPORTED,
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
ctx);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto done;
}
Start support for RTSP 2.0 This adds basic support for new 2.0 features, though the protocol is subposdely backward incompatible, most semantics are the sames. This commit adds: - features: * version negotiation * pipelined requests support * Media-Properties support * Accept-Ranges support - APIs: * gst_rtsp_media_seekable The RTSP methods that have been removed when using 2.0 now return BAD_REQUEST. https://bugzilla.gnome.org/show_bug.cgi?id=781446
2017-04-22 12:26:07 +00:00
invalid_command_for_version:
{
GST_ERROR ("client %p: invalid command for version", client);
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
goto done;
}
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
bad_request:
{
GST_ERROR ("client %p: bad request", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto done;
}
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
no_pool:
Initial import
2008-10-09 12:29:12 +00:00
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no pool configured", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto done;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
}
session_not_found:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: session not found", client);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_SESSION_NOT_FOUND, ctx);
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto done;
Initial import
2008-10-09 12:29:12 +00:00
}
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
not_authorized:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: not allowed", client);
client: Send setup reply once only If find_media() failed in handle_setup_request() two replies was sent. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=708819
2013-09-26 07:41:10 +00:00
/* error reply is already sent */
client: improve debug and fix leaks Cleanup the uri and session when there is a bad request.
2012-11-27 11:24:21 +00:00
goto done;
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
}
rtsp-client: handle Require headers and respond with OPTION_NOT_SUPPORTED Servers must handle Require headers and must report a failure if they don't handle any of the Required options, see RFC 2326, section 12.32: https://tools.ietf.org/html/rfc2326#page-54 https://bugzilla.gnome.org/show_bug.cgi?id=729426
2014-05-02 18:59:23 +00:00
unsupported_requirement:
{
GST_ERROR ("client %p: Required option is not supported (%s)", client,
unsupported_reqs);
send_option_not_supported_response (client, ctx, unsupported_reqs);
g_free (unsupported_reqs);
goto done;
}
client: improve debug
2012-11-27 11:33:02 +00:00
not_implemented:
{
GST_ERROR ("client %p: method %d not implemented", client, method);
ClientState -> Context Rename the clientstate to context and put the code in a separate file.
2013-07-22 12:25:04 +00:00
send_generic_response (client, GST_RTSP_STS_NOT_IMPLEMENTED, ctx);
client: improve debug
2012-11-27 11:33:02 +00:00
goto done;
}
Initial import
2008-10-09 12:29:12 +00:00
}
Add handle-response signal for when we receive a GET_PARAMETER response
2013-08-08 14:57:42 +00:00
static void
handle_response (GstRTSPClient * client, GstRTSPMessage * response)
{
GstRTSPClientPrivate *priv = client->priv;
GstRTSPResult res;
GstRTSPSession *session = NULL;
GstRTSPContext sctx = { NULL }, *ctx;
gchar *sessid;
if (!(ctx = gst_rtsp_context_get_current ())) {
ctx = &sctx;
ctx->auth = priv->auth;
gst_rtsp_context_push_current (ctx);
}
ctx->conn = priv->connection;
ctx->client = client;
ctx->request = NULL;
ctx->uri = NULL;
ctx->method = GST_RTSP_INVALID;
ctx->response = response;
if (gst_debug_category_get_threshold (rtsp_client_debug) >= GST_LEVEL_LOG) {
gst_rtsp_message_dump (response);
}
GST_INFO ("client %p: received a response", client);
/* get the session if there is any */
res =
gst_rtsp_message_get_header (response, GST_RTSP_HDR_SESSION, &sessid, 0);
if (res == GST_RTSP_OK) {
if (priv->session_pool == NULL)
goto no_pool;
/* we had a session in the request, find it again */
if (!(session = gst_rtsp_session_pool_find (priv->session_pool, sessid)))
goto session_not_found;
/* we add the session to the client list of watched sessions. When a session
* disappears because it times out, we will be notified. If all sessions are
* gone, we will close the connection */
client_watch_session (client, session);
}
ctx->session = session;
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_HANDLE_RESPONSE],
0, ctx);
done:
if (ctx == &sctx)
gst_rtsp_context_pop_current (ctx);
if (session)
g_object_unref (session);
return;
no_pool:
{
GST_ERROR ("client %p: no pool configured", client);
goto done;
}
session_not_found:
{
GST_ERROR ("client %p: session not found", client);
goto done;
}
}
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
static void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
handle_data (GstRTSPClient * client, GstRTSPMessage * message)
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
GstRTSPResult res;
guint8 channel;
guint8 *data;
guint size;
GstBuffer *buffer;
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
GstRTSPStreamTransport *trans;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
/* find the stream for this message */
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
res = gst_rtsp_message_parse_data (message, &channel);
if (res != GST_RTSP_OK)
return;
rtsp-client: Drop trailing \0 of RTSP DATA messages We add a trailing \0 in GstRTSPConnection to make parsing of string message bodies easier (e.g. the SDP from DESCRIBE) but for actual data this means we have to drop it or otherwise create invalid data.
2015-01-16 19:04:01 +00:00
gst_rtsp_message_get_body (message, &data, &size);
if (size < 2)
goto invalid_length;
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
gst_rtsp_message_steal_body (message, &data, &size);
rtsp-client: log interleaved data received
2015-01-19 20:35:15 +00:00
/* Strip trailing \0 (which GstRTSPConnection adds) */
--size;
buffer = gst_buffer_new_wrapped (data, size);
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
rtsp-client: Make old compilers happy rtsp-client.c:2553:50: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] Just in case that guint8 doesn't fit in a pointer. Just in case ...
2014-09-22 07:30:39 +00:00
trans =
g_hash_table_lookup (priv->transports, GINT_TO_POINTER ((gint) channel));
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
if (trans) {
rtsp-client: Place netaddress meta on packets received via TCP This allows us to later map signals from rtpbin/rtpsource back to the corresponding stream transport, and allows to do keep-alive based on RTCP packets in case of TCP media transport. https://bugzilla.gnome.org/show_bug.cgi?id=789646
2018-02-28 19:12:43 +00:00
GSocketAddress *addr;
/* Only create the socket address once for the transport, we don't really
* want to do that for every single packet.
*
* The netaddress meta is later used by the RTP stack to know where
* packets came from and allows us to match it again to a stream transport
*
* In theory we could use the remote socket address of the RTSP connection
* here, but this would fail with a custom configure_client_transport()
* implementation.
*/
if (!(addr =
g_object_get_data (G_OBJECT (trans), "rtsp-client.remote-addr"))) {
const GstRTSPTransport *tr;
GInetAddress *iaddr;
tr = gst_rtsp_stream_transport_get_transport (trans);
iaddr = g_inet_address_new_from_string (tr->destination);
if (iaddr) {
addr = g_inet_socket_address_new (iaddr, tr->client_port.min);
g_object_unref (iaddr);
g_object_set_data_full (G_OBJECT (trans), "rtsp-client.remote-addr",
addr, (GDestroyNotify) g_object_unref);
}
}
if (addr) {
gst_buffer_add_net_address_meta (buffer, addr);
}
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
/* dispatch to the stream based on the channel number */
rtsp-client: log interleaved data received
2015-01-19 20:35:15 +00:00
GST_LOG_OBJECT (client, "%u bytes of data on channel %u", size, channel);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
gst_rtsp_stream_transport_recv_data (trans, channel, buffer);
} else {
rtsp-client: log interleaved data received
2015-01-19 20:35:15 +00:00
GST_DEBUG_OBJECT (client, "received %u bytes of data for "
"unknown channel %u", size, channel);
server: use appsink and appsrc with the API Use the appsink/appsrc API instead of the signals for higher performance.
2009-04-14 21:38:58 +00:00
gst_buffer_unref (buffer);
client: simplify session transport handling link/unlink of the transport in a session was done to keep track of all TCP transports and to send RTP/RTCP data to the streams. We can simplify that by putting all the TCP transports in a hashtable indexed with the channel number. We also don't need to link/unlink the transports when we pause/resume the streams. The same effect is already achieved when we pause/play the media. Indeed, when we pause the media, the transport is removed from the media and the callbacks will not be called anymore. See https://bugzilla.gnome.org/show_bug.cgi?id=736041
2014-09-09 16:11:39 +00:00
}
rtsp-client: fix unintentional fallthrough to debug warning when receiving interleaved data
2015-01-19 20:18:20 +00:00
return;
/* ERRORS */
rtsp-client: Drop trailing \0 of RTSP DATA messages We add a trailing \0 in GstRTSPConnection to make parsing of string message bodies easier (e.g. the SDP from DESCRIBE) but for actual data this means we have to drop it or otherwise create invalid data.
2015-01-16 19:04:01 +00:00
invalid_length:
{
GST_DEBUG ("client %p: Short message received, ignoring", client);
return;
}
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
}
Initial import
2008-10-09 12:29:12 +00:00
/**
* gst_rtsp_client_set_session_pool:
* @client: a #GstRTSPClient
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* @pool: (transfer none) (nullable): a #GstRTSPSessionPool
Initial import
2008-10-09 12:29:12 +00:00
*
* Set @pool as the sessionpool for @client which it will use to find
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
* or allocate sessions. the sessionpool is usually inherited from the server
* that created the client but can be overridden later.
Initial import
2008-10-09 12:29:12 +00:00
*/
void
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
gst_rtsp_client_set_session_pool (GstRTSPClient * client,
GstRTSPSessionPool * pool)
Initial import
2008-10-09 12:29:12 +00:00
{
GstRTSPSessionPool *old;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
Initial import
2008-10-09 12:29:12 +00:00
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (pool)
g_object_ref (pool);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_lock (&priv->lock);
old = priv->session_pool;
priv->session_pool = pool;
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
if (priv->session_removed_id) {
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
g_signal_handler_disconnect (old, priv->session_removed_id);
priv->session_removed_id = 0;
client: keep ref to client for the session removed handler This extra ref will be dropped when all client sessions have been removed. A session is removed when a client sends teardown, closes its endpoint of the TCP connection or the sessions expires. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-09 13:01:31 +00:00
}
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_unlock (&priv->lock);
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
Client: keep a ref to the session Don't just keep a weak ref to the session objects but use a hard ref. We will be notified when a session is removed from the pool (expired) with the new session-removed signal. Don't automatically close the RTSP connection when all the sessions of a client are removed, a client can continue to operate and it can create a new session if it wants. If you want to remove the client from the server, you have to use gst_rtsp_server_client_filter() now. Based on patch from Ognyan Tonchev <ognyan.tonchev at axis.com> See https://bugzilla.gnome.org/show_bug.cgi?id=732226
2014-07-01 10:13:47 +00:00
/* FIXME, should remove all sessions from the old pool for this client */
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (old)
g_object_unref (old);
Initial import
2008-10-09 12:29:12 +00:00
}
/**
* gst_rtsp_client_get_session_pool:
* @client: a #GstRTSPClient
*
* Get the #GstRTSPSessionPool object that @client uses to manage its sessions.
*
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* Returns: (transfer full) (nullable): a #GstRTSPSessionPool, unref after usage.
Initial import
2008-10-09 12:29:12 +00:00
*/
GstRTSPSessionPool *
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
gst_rtsp_client_get_session_pool (GstRTSPClient * client)
Initial import
2008-10-09 12:29:12 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
Initial import
2008-10-09 12:29:12 +00:00
GstRTSPSessionPool *result;
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
g_mutex_lock (&priv->lock);
if ((result = priv->session_pool))
Initial import
2008-10-09 12:29:12 +00:00
g_object_ref (result);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_unlock (&priv->lock);
Initial import
2008-10-09 12:29:12 +00:00
return result;
client: add support for setting the server. Add support for keeping a ref to the server that started this client connection.
2011-01-12 09:42:52 +00:00
}
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
/**
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
* gst_rtsp_client_set_mount_points:
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
* @client: a #GstRTSPClient
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* @mounts: (transfer none) (nullable): a #GstRTSPMountPoints
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
*
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
* Set @mounts as the mount points for @client which it will use to map urls
* to media streams. These mount points are usually inherited from the server that
Make the server handle arbitrary pipelines Make GstMediaFactory an object that can instantiate GstMediaBin objects. The GstMediaBin object has a handle to a bin with elements and to a list of GstMediaStream objects that this bin produces. Add GstMediaMapper that can map url mountpoints to GstMediaFactory objects along with methods to register and remove those mappings. Add methods and a property to GstRTSPServer to manage the GstMediaMapper object used by the server instance. Modify the example application so that it shows how to create custom pipelines attached to a specific mount point. Various misc cleanps.
2009-01-22 14:33:29 +00:00
* created the client but can be overriden later.
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
*/
void
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
gst_rtsp_client_set_mount_points (GstRTSPClient * client,
GstRTSPMountPoints * mounts)
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
GstRTSPMountPoints *old;
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (mounts)
g_object_ref (mounts);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_lock (&priv->lock);
old = priv->mount_points;
priv->mount_points = mounts;
g_mutex_unlock (&priv->lock);
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (old)
g_object_unref (old);
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
}
/**
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
* gst_rtsp_client_get_mount_points:
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
* @client: a #GstRTSPClient
*
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
* Get the #GstRTSPMountPoints object that @client uses to manage its sessions.
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
*
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* Returns: (transfer full) (nullable): a #GstRTSPMountPoints, unref after usage.
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
*/
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
GstRTSPMountPoints *
gst_rtsp_client_get_mount_points (GstRTSPClient * client)
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
MediaMapping -> MountPoints Describes better what the object manages.
2012-11-26 11:34:05 +00:00
GstRTSPMountPoints *result;
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
g_mutex_lock (&priv->lock);
if ((result = priv->mount_points))
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
g_object_ref (result);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_unlock (&priv->lock);
Allow setting a custom media factory for a client.
2009-01-20 18:46:21 +00:00
return result;
}
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
/**
* gst_rtsp_client_set_content_length_limit:
* @client: a #GstRTSPClient
* @limit: Content-Length limit
*
* Configure @client to use the specified Content-Length limit.
*
* Define an appropriate request size limit and reject requests exceeding the
* limit with response status 413 Request Entity Too Large
*
* Since: 1.18
*/
void
gst_rtsp_client_set_content_length_limit (GstRTSPClient * client, guint limit)
{
GstRTSPClientPrivate *priv;
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
priv = client->priv;
g_mutex_lock (&priv->lock);
priv->content_length_limit = limit;
g_mutex_unlock (&priv->lock);
}
/**
* gst_rtsp_client_get_content_length_limit:
* @client: a #GstRTSPClient
*
* Get the Content-Length limit of @client.
*
* Returns: the Content-Length limit.
*
* Since: 1.18
*/
guint
gst_rtsp_client_get_content_length_limit (GstRTSPClient * client)
{
GstRTSPClientPrivate *priv;
glong content_length_limit;
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), -1);
priv = client->priv;
g_mutex_lock (&priv->lock);
content_length_limit = priv->content_length_limit;
g_mutex_unlock (&priv->lock);
return content_length_limit;
}
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
/**
* gst_rtsp_client_set_auth:
* @client: a #GstRTSPClient
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* @auth: (transfer none) (nullable): a #GstRTSPAuth
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
*
* configure @auth to be used as the authentication manager of @client.
*/
void
gst_rtsp_client_set_auth (GstRTSPClient * client, GstRTSPAuth * auth)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
GstRTSPAuth *old;
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (auth)
g_object_ref (auth);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_lock (&priv->lock);
old = priv->auth;
priv->auth = auth;
g_mutex_unlock (&priv->lock);
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
client: add locking Also add some g_return_if()
2012-11-26 15:30:16 +00:00
if (old)
g_object_unref (old);
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
}
/**
* gst_rtsp_client_get_auth:
* @client: a #GstRTSPClient
*
* Get the #GstRTSPAuth used as the authentication manager of @client.
*
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* Returns: (transfer full) (nullable): the #GstRTSPAuth of @client.
* g_object_unref() after usage.
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
*/
GstRTSPAuth *
gst_rtsp_client_get_auth (GstRTSPClient * client)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
GstRTSPAuth *result;
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
g_mutex_lock (&priv->lock);
if ((result = priv->auth))
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
g_object_ref (result);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_mutex_unlock (&priv->lock);
auth: add authentication object Add an object that can check the authorization of requests. Implement basic authentication. Add example authentication to test-video
2011-01-11 23:17:54 +00:00
return result;
}
client: add method to configure thread pool
2013-07-10 15:06:36 +00:00
/**
* gst_rtsp_client_set_thread_pool:
* @client: a #GstRTSPClient
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* @pool: (transfer none) (nullable): a #GstRTSPThreadPool
client: add method to configure thread pool
2013-07-10 15:06:36 +00:00
*
* configure @pool to be used as the thread pool of @client.
*/
void
gst_rtsp_client_set_thread_pool (GstRTSPClient * client,
GstRTSPThreadPool * pool)
{
GstRTSPClientPrivate *priv;
GstRTSPThreadPool *old;
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
priv = client->priv;
if (pool)
g_object_ref (pool);
g_mutex_lock (&priv->lock);
old = priv->thread_pool;
priv->thread_pool = pool;
g_mutex_unlock (&priv->lock);
if (old)
g_object_unref (old);
}
/**
* gst_rtsp_client_get_thread_pool:
* @client: a #GstRTSPClient
*
* Get the #GstRTSPThreadPool used as the thread pool of @client.
*
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* Returns: (transfer full) (nullable): the #GstRTSPThreadPool of @client. g_object_unref() after
client: add method to configure thread pool
2013-07-10 15:06:36 +00:00
* usage.
*/
GstRTSPThreadPool *
gst_rtsp_client_get_thread_pool (GstRTSPClient * client)
{
GstRTSPClientPrivate *priv;
GstRTSPThreadPool *result;
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
priv = client->priv;
g_mutex_lock (&priv->lock);
if ((result = priv->thread_pool))
g_object_ref (result);
g_mutex_unlock (&priv->lock);
return result;
}
server: refactor connection setup Let the server accept the socket connection and construct a GstRTSPConnection from it. Remove the code from the client and let the client only deal with a fully configure GstRTSPConnection object. We will need this later when the server will configure the connection for TLS.
2013-05-30 05:18:22 +00:00
/**
* gst_rtsp_client_set_connection:
* @client: a #GstRTSPClient
* @conn: (transfer full): a #GstRTSPConnection
*
* Set the #GstRTSPConnection of @client. This function takes ownership of
* @conn.
*
* Returns: %TRUE on success.
*/
gboolean
gst_rtsp_client_set_connection (GstRTSPClient * client,
GstRTSPConnection * conn)
{
GstRTSPClientPrivate *priv;
GSocket *read_socket;
GSocketAddress *address;
GstRTSPUrl *url;
GError *error = NULL;
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), FALSE);
g_return_val_if_fail (conn != NULL, FALSE);
priv = client->priv;
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
gst_rtsp_connection_set_content_length_limit (conn,
priv->content_length_limit);
server: refactor connection setup Let the server accept the socket connection and construct a GstRTSPConnection from it. Remove the code from the client and let the client only deal with a fully configure GstRTSPConnection object. We will need this later when the server will configure the connection for TLS.
2013-05-30 05:18:22 +00:00
read_socket = gst_rtsp_connection_get_read_socket (conn);
use local address, not remote, in SDP See https://bugzilla.gnome.org/show_bug.cgi?id=702063
2013-06-12 08:56:16 +00:00
if (!(address = g_socket_get_local_address (read_socket, &error)))
server: refactor connection setup Let the server accept the socket connection and construct a GstRTSPConnection from it. Remove the code from the client and let the client only deal with a fully configure GstRTSPConnection object. We will need this later when the server will configure the connection for TLS.
2013-05-30 05:18:22 +00:00
goto no_address;
g_free (priv->server_ip);
/* keep the original ip that the client connected to */
if (G_IS_INET_SOCKET_ADDRESS (address)) {
GInetAddress *iaddr;
iaddr = g_inet_socket_address_get_address (G_INET_SOCKET_ADDRESS (address));
/* socket might be ipv6 but adress still ipv4 */
priv->is_ipv6 = g_inet_address_get_family (iaddr) == G_SOCKET_FAMILY_IPV6;
priv->server_ip = g_inet_address_to_string (iaddr);
g_object_unref (address);
} else {
priv->is_ipv6 = g_socket_get_family (read_socket) == G_SOCKET_FAMILY_IPV6;
priv->server_ip = g_strdup ("unknown");
}
GST_INFO ("client %p connected to server ip %s, ipv6 = %d", client,
priv->server_ip, priv->is_ipv6);
url = gst_rtsp_connection_get_url (conn);
GST_INFO ("added new client %p ip %s:%d", client, url->host, url->port);
priv->connection = conn;
return TRUE;
/* ERRORS */
no_address:
{
client: fix comment
2013-07-01 13:18:43 +00:00
GST_ERROR ("could not get local address %s", error->message);
server: refactor connection setup Let the server accept the socket connection and construct a GstRTSPConnection from it. Remove the code from the client and let the client only deal with a fully configure GstRTSPConnection object. We will need this later when the server will configure the connection for TLS.
2013-05-30 05:18:22 +00:00
g_error_free (error);
return FALSE;
}
}
client: expose connection Fixes https://bugzilla.gnome.org/show_bug.cgi?id=697546
2013-04-15 10:17:34 +00:00
/**
* gst_rtsp_client_get_connection:
* @client: a #GstRTSPClient
*
* Get the #GstRTSPConnection of @client.
*
All around: add annotations and API guards
2018-02-09 23:07:25 +00:00
* Returns: (transfer none) (nullable): the #GstRTSPConnection of @client.
client: expose connection Fixes https://bugzilla.gnome.org/show_bug.cgi?id=697546
2013-04-15 10:17:34 +00:00
* The connection object returned remains valid until the client is freed.
*/
GstRTSPConnection *
gst_rtsp_client_get_connection (GstRTSPClient * client)
{
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
return client->priv->connection;
}
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
/**
* gst_rtsp_client_set_send_func:
* @client: a #GstRTSPClient
gobject-introspection: Add annotations to support language bindings In addition a few cosmetic changes: * Adjust the order of arguments * Fix typo: occured -> occurred * Fix indentation after Return:-clauses Fixes https://bugzilla.gnome.org/show_bug.cgi?id=726941
2014-03-23 20:24:48 +00:00
* @func: (scope notified): a #GstRTSPClientSendFunc
* @user_data: (closure): user data passed to @func
* @notify: (allow-none): called when @user_data is no longer in use
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
*
* Set @func as the callback that will be called when a new message needs to be
* sent to the client. @user_data is passed to @func and @notify is called when
* @user_data is no longer in use.
docs: improve docs
2013-07-11 14:57:14 +00:00
*
* By default, the client will send the messages on the #GstRTSPConnection that
* was configured with gst_rtsp_client_attach() was called.
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
*
* It is only allowed to set either a `send_func` or a `send_messages_func`
* but not both at the same time.
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
*/
void
gst_rtsp_client_set_send_func (GstRTSPClient * client,
GstRTSPClientSendFunc func, gpointer user_data, GDestroyNotify notify)
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
client: small cleanup
2012-11-26 16:35:51 +00:00
GDestroyNotify old_notify;
client: call destroy without the lock
2012-11-26 16:28:29 +00:00
gpointer old_data;
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_lock (&priv->send_lock);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
g_assert (func == NULL || priv->send_messages_func == NULL);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->send_func = func;
old_notify = priv->send_notify;
old_data = priv->send_data;
priv->send_notify = notify;
priv->send_data = user_data;
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
g_mutex_unlock (&priv->send_lock);
client: call destroy without the lock
2012-11-26 16:28:29 +00:00
client: small cleanup
2012-11-26 16:35:51 +00:00
if (old_notify)
old_notify (old_data);
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
}
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
/**
* gst_rtsp_client_set_send_messages_func:
* @client: a #GstRTSPClient
* @func: (scope notified): a #GstRTSPClientSendMessagesFunc
* @user_data: (closure): user data passed to @func
* @notify: (allow-none): called when @user_data is no longer in use
*
* Set @func as the callback that will be called when new messages needs to be
* sent to the client. @user_data is passed to @func and @notify is called when
* @user_data is no longer in use.
*
* By default, the client will send the messages on the #GstRTSPConnection that
* was configured with gst_rtsp_client_attach() was called.
*
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
* It is only allowed to set either a `send_func` or a `send_messages_func`
* but not both at the same time.
*
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
* Since: 1.16
*/
void
gst_rtsp_client_set_send_messages_func (GstRTSPClient * client,
GstRTSPClientSendMessagesFunc func, gpointer user_data,
GDestroyNotify notify)
{
GstRTSPClientPrivate *priv;
GDestroyNotify old_notify;
gpointer old_data;
g_return_if_fail (GST_IS_RTSP_CLIENT (client));
priv = client->priv;
g_mutex_lock (&priv->send_lock);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
g_assert (func == NULL || priv->send_func == NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
priv->send_messages_func = func;
old_notify = priv->send_messages_notify;
old_data = priv->send_messages_data;
priv->send_messages_notify = notify;
priv->send_messages_data = user_data;
g_mutex_unlock (&priv->send_lock);
if (old_notify)
old_notify (old_data);
}
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
/**
* gst_rtsp_client_handle_message:
* @client: a #GstRTSPClient
gobject-introspection: Add annotations to support language bindings In addition a few cosmetic changes: * Adjust the order of arguments * Fix typo: occured -> occurred * Fix indentation after Return:-clauses Fixes https://bugzilla.gnome.org/show_bug.cgi?id=726941
2014-03-23 20:24:48 +00:00
* @message: (transfer none): an #GstRTSPMessage
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
*
* Let the client handle @message.
*
* Returns: a #GstRTSPResult.
*/
GstRTSPResult
gst_rtsp_client_handle_message (GstRTSPClient * client,
GstRTSPMessage * message)
{
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), GST_RTSP_EINVAL);
g_return_val_if_fail (message != NULL, GST_RTSP_EINVAL);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
switch (message->type) {
case GST_RTSP_MESSAGE_REQUEST:
handle_request (client, message);
break;
case GST_RTSP_MESSAGE_RESPONSE:
Add handle-response signal for when we receive a GET_PARAMETER response
2013-08-08 14:57:42 +00:00
handle_response (client, message);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
break;
case GST_RTSP_MESSAGE_DATA:
Add TCP transports Use appsrc and appsink to send and receive RTP/RTCP packets in the TCP connection.
2009-03-11 15:45:12 +00:00
handle_data (client, message);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
break;
default:
break;
}
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
return GST_RTSP_OK;
}
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
/**
client: allow for sending any message, not only requests Change the _send_request() method to _send_message() so that we can both send requests and replies.
2013-07-12 13:13:48 +00:00
* gst_rtsp_client_send_message:
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
* @client: a #GstRTSPClient
introspection: add missing allow-none annotations https://bugzilla.gnome.org/show_bug.cgi?id=730952
2014-06-11 23:42:08 +00:00
* @session: (allow-none) (transfer none): a #GstRTSPSession to send
* the message to or %NULL
gobject-introspection: Add annotations to support language bindings In addition a few cosmetic changes: * Adjust the order of arguments * Fix typo: occured -> occurred * Fix indentation after Return:-clauses Fixes https://bugzilla.gnome.org/show_bug.cgi?id=726941
2014-03-23 20:24:48 +00:00
* @message: (transfer none): The #GstRTSPMessage to send
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
*
client: allow for sending any message, not only requests Change the _send_request() method to _send_message() so that we can both send requests and replies.
2013-07-12 13:13:48 +00:00
* Send a message message to the remote end. @message must be a
* #GST_RTSP_MESSAGE_REQUEST or a #GST_RTSP_MESSAGE_RESPONSE.
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
*/
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
GstRTSPResult
client: allow for sending any message, not only requests Change the _send_request() method to _send_message() so that we can both send requests and replies.
2013-07-12 13:13:48 +00:00
gst_rtsp_client_send_message (GstRTSPClient * client, GstRTSPSession * session,
GstRTSPMessage * message)
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
{
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
GstRTSPContext sctx = { NULL }
, *ctx;
GstRTSPClientPrivate *priv;
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), GST_RTSP_EINVAL);
client: allow for sending any message, not only requests Change the _send_request() method to _send_message() so that we can both send requests and replies.
2013-07-12 13:13:48 +00:00
g_return_val_if_fail (message != NULL, GST_RTSP_EINVAL);
g_return_val_if_fail (message->type == GST_RTSP_MESSAGE_REQUEST ||
message->type == GST_RTSP_MESSAGE_RESPONSE, GST_RTSP_EINVAL);
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
client: pass context to send_message Pass the current context to send_message, we will need it later.
2014-05-01 04:07:08 +00:00
priv = client->priv;
if (!(ctx = gst_rtsp_context_get_current ())) {
ctx = &sctx;
ctx->auth = priv->auth;
gst_rtsp_context_push_current (ctx);
}
ctx->conn = priv->connection;
ctx->client = client;
ctx->session = session;
send_message (client, ctx, message, FALSE);
if (ctx == &sctx)
gst_rtsp_context_pop_current (ctx);
client: clean some variables Clean some variables and add some guards to _send_request()
2013-06-25 13:09:13 +00:00
return GST_RTSP_OK;
Add gst_rtsp_client_send_request API This makes it possible to send arbitrary messages to a client, such as SET_PARAMETER or GET_PARAMETER
2013-06-10 21:32:12 +00:00
}
rtsp-client: add gst_rtsp_client_get_stream_transport() This will be used in the onvif tests in order to validate the data transmitted over TCP: for streaming to continue after a data message has been provided to client->send_func, the client is responsible for marking the message as sent on the relevant stream transport.
2018-11-05 14:34:20 +00:00
/**
* gst_rtsp_client_get_stream_transport:
*
* This is useful when providing a send function through
* gst_rtsp_client_set_send_func() when doing RTSP over TCP:
* the send function must call gst_rtsp_stream_transport_message_sent ()
* on the appropriate transport when data has been received for streaming
* to continue.
*
* Returns: (transfer none) (nullable): the #GstRTSPStreamTransport associated with @channel.
*
* Since: 1.18
*/
GstRTSPStreamTransport *
gst_rtsp_client_get_stream_transport (GstRTSPClient * self, guint8 channel)
{
return g_hash_table_lookup (self->priv->transports,
GINT_TO_POINTER ((gint) channel));
}
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
static gboolean
do_send_messages (GstRTSPClient * client, GstRTSPMessage * messages,
guint n_messages, gboolean close, gpointer user_data)
{
GstRTSPClientPrivate *priv = client->priv;
guint id = 0;
GstRTSPResult ret;
guint i;
/* send the message */
rtsp-client: If the TEARDOWN response can be sent directly, directly close the client Instead of closing it never at all. Previously there was only code that closed the client asynchronously if sending the response happened asynchrously at a later time. Thanks to Christian M for debugging this issue. Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/issues/102 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/114>
2020-04-27 10:49:55 +00:00
if (close)
GST_INFO ("client %p: sending close message", client);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
ret = gst_rtsp_watch_send_messages (priv->watch, messages, n_messages, &id);
if (ret != GST_RTSP_OK)
goto error;
for (i = 0; i < n_messages; i++) {
if (gst_rtsp_message_get_type (&messages[i]) == GST_RTSP_MESSAGE_DATA) {
guint8 channel = 0;
GstRTSPResult r;
/* We assume that all data messages in the list are for the
* same channel */
r = gst_rtsp_message_parse_data (&messages[i], &channel);
if (r != GST_RTSP_OK) {
ret = r;
goto error;
}
/* check if the message has been queued for transmission in watch */
if (id) {
/* store the seq number so we can wait until it has been sent */
GST_DEBUG_OBJECT (client, "wait for message %d, channel %d", id,
channel);
set_data_seq (client, channel, id);
} else {
GstRTSPStreamTransport *trans;
trans =
g_hash_table_lookup (priv->transports,
GINT_TO_POINTER ((gint) channel));
if (trans) {
GST_DEBUG_OBJECT (client, "emit 'message-sent' signal");
g_mutex_unlock (&priv->send_lock);
gst_rtsp_stream_transport_message_sent (trans);
g_mutex_lock (&priv->send_lock);
}
}
break;
}
}
return ret == GST_RTSP_OK;
/* ERRORS */
error:
{
GST_DEBUG_OBJECT (client, "got error %d", ret);
return FALSE;
}
}
client: make the client usable without a socket Make a method to let the client handle a message and a callback when the client wants us to send a response message back. This makes it possible to also use the client object without the sockets, which should make it easier to test.
2012-11-26 16:20:39 +00:00
static GstRTSPResult
message_received (GstRTSPWatch * watch, GstRTSPMessage * message,
gpointer user_data)
{
return gst_rtsp_client_handle_message (GST_RTSP_CLIENT (user_data), message);
}
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
static GstRTSPResult
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
message_sent (GstRTSPWatch * watch, guint cseq, gpointer user_data)
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
{
client: small cleanup
2012-11-26 15:45:04 +00:00
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
Limit queued TCP data messages to one per stream Before, the watch backlog size in GstRTSPClient was changed dynamically between unlimited and a fixed size, trying to avoid both unlimited memory usage and deadlocks while waiting for place in the queue. (Some of the deadlocks were described in a long comment in handle_request().) In the previous commit, we changed to a fixed backlog size of 100. This is possible, because we now handle RTP/RTCP data messages differently from RTSP request/response messages. The data messages are messages tunneled over TCP. We allow at most one queued data message per stream in GstRTSPClient at a time, and successfully sent data messages are acked by sending a "message-sent" callback from the GstStreamTransport. Until that ack comes, the GstRTSPStream does not call pull_sample() on its appsink, and therefore the streaming thread in the pipeline will not be blocked inside GstRTSPClient, waiting for a place in the queue. pull_sample() is called when we have both an ack and a "new-sample" signal from the appsink. Then, we know there is a buffer to write. RTSP request/response messages are not acked in the same way as data messages. The rest of the 100 places in the queue are used for them. If the queue becomes full of request/response messages, we return an error and close the connection to the client. Change-Id: I275310bc90a219ceb2473c098261acc78be84c97
2018-06-28 09:22:21 +00:00
GstRTSPStreamTransport *trans = NULL;
guint8 channel = 0;
g_mutex_lock (&priv->send_lock);
if (get_data_channel (client, cseq, &channel)) {
trans = g_hash_table_lookup (priv->transports, GINT_TO_POINTER (channel));
set_data_seq (client, channel, 0);
}
g_mutex_unlock (&priv->send_lock);
if (trans) {
GST_DEBUG_OBJECT (client, "emit 'message-sent' signal");
gst_rtsp_stream_transport_message_sent (trans);
}
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
return GST_RTSP_OK;
}
static GstRTSPResult
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
closed (GstRTSPWatch * watch, gpointer user_data)
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
{
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
const gchar *tunnelid;
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
Use GStreamer's debugging subsystem
2009-11-21 18:20:23 +00:00
GST_INFO ("client %p: connection closed", client);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
if ((tunnelid = gst_rtsp_connection_get_tunnelid (priv->connection))) {
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_lock (&tunnels_lock);
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
/* remove from tunnelids */
Only free the pending tunnel if there is one --
2009-03-04 15:33:21 +00:00
g_hash_table_remove (tunnels, tunnelid);
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_unlock (&tunnels_lock);
Only free the pending tunnel if there is one --
2009-03-04 15:33:21 +00:00
}
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: Add drop-backlog property When we have too many messages queued for a client (currently hardcoded to 100) we overflow and drop the messages. Add a drop-backlog property to control this behaviour. Setting this property to FALSE will retry to send the messages to the client by waiting for more room in the backlog. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725898
2014-04-01 11:04:21 +00:00
gst_rtsp_watch_set_flushing (watch, TRUE);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_lock (&priv->watch_lock);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_client_set_send_messages_func (client, NULL, NULL, NULL);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_unlock (&priv->watch_lock);
client: make sure the watch exists while sending data Protect the send_func with a lock. This allows us to wait for sending to complete before changing the send_func and user_data. We add an extra ref to the watch to make sure that it remains valid during sending. When closing the connection, set the send_func to NULL Fixes https://bugzilla.gnome.org/show_bug.cgi?id=692433
2013-01-28 09:31:50 +00:00
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
return GST_RTSP_OK;
}
static GstRTSPResult
rtsp: use RTCP to keep the session alive Use the RTCP rtcp-from stats field to find the associated session and use this to keep the session alive.
2009-05-26 17:01:10 +00:00
error (GstRTSPWatch * watch, GstRTSPResult result, gpointer user_data)
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
{
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
gchar *str;
str = gst_rtsp_strresult (result);
Use GStreamer's debugging subsystem
2009-11-21 18:20:23 +00:00
GST_INFO ("client %p: received an error %s", client, str);
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
g_free (str);
return GST_RTSP_OK;
}
rtsp-client: implement error_full Implement error_full to avoid some segfaults when the rtspconnection calls it. See #608245
2010-01-27 17:38:27 +00:00
static GstRTSPResult
client: use right size for malloc
2010-03-16 17:33:23 +00:00
error_full (GstRTSPWatch * watch, GstRTSPResult result,
GstRTSPMessage * message, guint id, gpointer user_data)
rtsp-client: implement error_full Implement error_full to avoid some segfaults when the rtspconnection calls it. See #608245
2010-01-27 17:38:27 +00:00
{
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
gchar *str;
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
GstRTSPContext sctx = { NULL }, *ctx;
GstRTSPClientPrivate *priv;
GstRTSPMessage response = { 0 };
priv = client->priv;
if (!(ctx = gst_rtsp_context_get_current ())) {
ctx = &sctx;
ctx->auth = priv->auth;
gst_rtsp_context_push_current (ctx);
}
ctx->conn = priv->connection;
ctx->client = client;
ctx->request = message;
ctx->method = GST_RTSP_INVALID;
ctx->response = &response;
rtsp-client: implement error_full Implement error_full to avoid some segfaults when the rtspconnection calls it. See #608245
2010-01-27 17:38:27 +00:00
rtsp-client: Handle Content-Length limitation Add functionality to limit the Content-Length. API addition, Enhancement. Define an appropriate request size limit and reject requests exceeding the limit with response status 413 Request Entity Too Large Related to !182
2019-03-27 15:21:03 +00:00
/* only return error response if it is a request */
if (!message || message->type != GST_RTSP_MESSAGE_REQUEST)
goto done;
if (result == GST_RTSP_ENOMEM) {
send_generic_response (client, GST_RTSP_STS_REQUEST_ENTITY_TOO_LARGE, ctx);
goto done;
}
if (result == GST_RTSP_EPARSE) {
send_generic_response (client, GST_RTSP_STS_BAD_REQUEST, ctx);
goto done;
}
done:
if (ctx == &sctx)
gst_rtsp_context_pop_current (ctx);
rtsp-client: implement error_full Implement error_full to avoid some segfaults when the rtspconnection calls it. See #608245
2010-01-27 17:38:27 +00:00
str = gst_rtsp_strresult (result);
client: use right size for malloc
2010-03-16 17:33:23 +00:00
GST_INFO
client: improve debug
2013-05-30 08:52:46 +00:00
("client %p: error when handling message %p with id %d: %s",
client, message, id, str);
rtsp-client: implement error_full Implement error_full to avoid some segfaults when the rtspconnection calls it. See #608245
2010-01-27 17:38:27 +00:00
g_free (str);
return GST_RTSP_OK;
}
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
static gboolean
remember_tunnel (GstRTSPClient * client)
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
const gchar *tunnelid;
/* store client in the pending tunnels */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
tunnelid = gst_rtsp_connection_get_tunnelid (priv->connection);
client: don't crash when tunnelid is missing When a clients tries to open an HTTP tunnel but fails to provide a tunnelid, don't crash but return an error response to the client. Fixes #589489
2009-07-24 10:49:41 +00:00
if (tunnelid == NULL)
goto no_tunnelid;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
GST_INFO ("client %p: inserting tunnel session %s", client, tunnelid);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
/* we can't have two clients connecting with the same tunnelid */
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_lock (&tunnels_lock);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
if (g_hash_table_lookup (tunnels, tunnelid))
goto tunnel_existed;
g_hash_table_insert (tunnels, g_strdup (tunnelid), g_object_ref (client));
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_unlock (&tunnels_lock);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
return TRUE;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
/* ERRORS */
client: don't crash when tunnelid is missing When a clients tries to open an HTTP tunnel but fails to provide a tunnelid, don't crash but return an error response to the client. Fixes #589489
2009-07-24 10:49:41 +00:00
no_tunnelid:
{
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
GST_ERROR ("client %p: no tunnelid provided", client);
return FALSE;
client: don't crash when tunnelid is missing When a clients tries to open an HTTP tunnel but fails to provide a tunnelid, don't crash but return an error response to the client. Fixes #589489
2009-07-24 10:49:41 +00:00
}
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
tunnel_existed:
{
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_unlock (&tunnels_lock);
rtsp-server: Run gst-indent Since it wasn't using the upstream common previously, there was no indentation check before commiting.
2010-12-11 09:48:42 +00:00
GST_ERROR ("client %p: tunnel session %s already existed", client,
tunnelid);
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
return FALSE;
}
}
static GstRTSPResult
tunnel_lost (GstRTSPWatch * watch, gpointer user_data)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
GstRTSPClientPrivate *priv = client->priv;
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
client: log more errors
2012-11-26 12:37:20 +00:00
GST_WARNING ("client %p: tunnel lost (connection %p)", client,
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->connection);
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
/* ignore error, it'll only be a problem when the client does a POST again */
remember_tunnel (client);
return GST_RTSP_OK;
}
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
static GstRTSPStatusCode
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
handle_tunnel (GstRTSPClient * client)
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
GstRTSPClient *oclient;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *opriv;
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
const gchar *tunnelid;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
tunnelid = gst_rtsp_connection_get_tunnelid (priv->connection);
client: don't crash when tunnelid is missing When a clients tries to open an HTTP tunnel but fails to provide a tunnelid, don't crash but return an error response to the client. Fixes #589489
2009-07-24 10:49:41 +00:00
if (tunnelid == NULL)
goto no_tunnelid;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
/* check for previous tunnel */
rtsp-server: port to new thread API
2012-05-11 07:42:47 +00:00
g_mutex_lock (&tunnels_lock);
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
oclient = g_hash_table_lookup (tunnels, tunnelid);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
if (oclient == NULL) {
/* no previous tunnel, remember tunnel */
g_hash_table_insert (tunnels, g_strdup (tunnelid), g_object_ref (client));
g_mutex_unlock (&tunnels_lock);
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
GST_INFO ("client %p: no previous tunnel found, remembering tunnel (%p)",
client, priv->connection);
} else {
/* merge both tunnels into the first client */
/* remove the old client from the table. ref before because removing it will
* remove the ref to it. */
g_object_ref (oclient);
g_hash_table_remove (tunnels, tunnelid);
g_mutex_unlock (&tunnels_lock);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
opriv = oclient->priv;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_lock (&opriv->watch_lock);
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
if (opriv->watch == NULL)
goto tunnel_closed;
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
if (opriv->tstate == priv->tstate)
goto tunnel_duplicate_id;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
GST_INFO ("client %p: found previous tunnel %p (old %p, new %p)", client,
oclient, opriv->connection, priv->connection);
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
gst_rtsp_connection_do_tunnel (opriv->connection, priv->connection);
gst_rtsp_watch_reset (priv->watch);
gst_rtsp_watch_reset (opriv->watch);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_unlock (&opriv->watch_lock);
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
g_object_unref (oclient);
client: remove watch of the second client after http tunnel setup The second client will be freed after the HTTP tunnel has been set up. Make sure it's RTSP watch is never dispatched again. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=727488
2014-04-02 10:03:32 +00:00
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
/* the old client owns the tunnel now, the new one will be freed */
g_source_destroy ((GSource *) priv->watch);
priv->watch = NULL;
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_client_set_send_messages_func (client, NULL, NULL, NULL);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove (client);
rtsp-client: Clean up watch/watch context and related state consistently And assert that it was cleaned up properly before the client is finalized. If something is still around when the client is shut down then something went very wrong before. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:16:18 +00:00
}
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
return GST_RTSP_STS_OK;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
/* ERRORS */
client: don't crash when tunnelid is missing When a clients tries to open an HTTP tunnel but fails to provide a tunnelid, don't crash but return an error response to the client. Fixes #589489
2009-07-24 10:49:41 +00:00
no_tunnelid:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: no tunnelid provided", client);
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
return GST_RTSP_STS_SERVICE_UNAVAILABLE;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
}
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
tunnel_closed:
{
client: log more errors
2012-11-26 12:37:20 +00:00
GST_ERROR ("client %p: tunnel session %s was closed", client, tunnelid);
rtsp-client: Protect saved clients watch with a mutex Fixes a crash when close() is called while merging clients in handle_tunnel(). In that case close() would destroy the watch while it is still being used in handle_tunnel(). https://bugzilla.gnome.org/show_bug.cgi?id=735570
2014-08-28 11:35:15 +00:00
g_mutex_unlock (&opriv->watch_lock);
client: fix unlink on session timeouts When our session times out, make sure we unlink all streams in this session. Remove the tunnelid when closing the connection.
2010-04-06 13:45:56 +00:00
g_object_unref (oclient);
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
return GST_RTSP_STS_SERVICE_UNAVAILABLE;
}
tunnel_duplicate_id:
{
GST_ERROR ("client %p: tunnel session %s was duplicate", client, tunnelid);
g_mutex_unlock (&opriv->watch_lock);
g_object_unref (oclient);
return GST_RTSP_STS_BAD_REQUEST;
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
}
}
static GstRTSPStatusCode
tunnel_get (GstRTSPWatch * watch, gpointer user_data)
{
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
GST_INFO ("client %p: tunnel get (connection %p)", client,
client->priv->connection);
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
g_mutex_lock (&client->priv->lock);
client->priv->tstate = TUNNEL_STATE_GET;
g_mutex_unlock (&client->priv->lock);
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
return handle_tunnel (client);
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
}
static GstRTSPResult
tunnel_post (GstRTSPWatch * watch, gpointer user_data)
{
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
GST_INFO ("client %p: tunnel post (connection %p)", client,
client->priv->connection);
rtsp-client:Error handling when equal http session cookie There are some clients that are sending same session cookie on random basis. https://bugzilla.gnome.org/show_bug.cgi?id=753616
2018-02-15 12:26:16 +00:00
g_mutex_lock (&client->priv->lock);
client->priv->tstate = TUNNEL_STATE_POST;
g_mutex_unlock (&client->priv->lock);
if (handle_tunnel (client) != GST_RTSP_STS_OK)
rtsp-server: Fix compilation and compiler warnings
2012-04-13 13:27:22 +00:00
return GST_RTSP_ERROR;
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
return GST_RTSP_OK;
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
}
rtsp-client: vmethod for modifying tunnel GET response Add a vmethod tunnel_http_response where the response to the HTTP GET for tunneled connections can be modified. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725879
2014-03-06 12:52:02 +00:00
static GstRTSPResult
tunnel_http_response (GstRTSPWatch * watch, GstRTSPMessage * request,
GstRTSPMessage * response, gpointer user_data)
{
GstRTSPClientClass *klass;
GstRTSPClient *client = GST_RTSP_CLIENT (user_data);
klass = GST_RTSP_CLIENT_GET_CLASS (client);
if (klass->tunnel_http_response) {
klass->tunnel_http_response (client, request, response);
}
return GST_RTSP_OK;
}
Fix for channel -> watch rename in gstreamer Rename the RTSPChannel to RTSPWatch and remove an unused variable.
2009-02-19 14:53:50 +00:00
static GstRTSPWatchFuncs watch_funcs = {
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
message_received,
message_sent,
closed,
rtsp-server: Add support for tunneling Add support for tunneling over HTTP. Use new connection methods to retrieve the url. Dispatch messages based on the message type instead of blindly assuming it's always a request. Keep track of the watch id so that we can remove it later. Set the media pipeline to NULL before unreffing the pipeline.
2009-03-04 11:44:01 +00:00
error,
client: support for POST before GET when setting up a tunnel
2014-04-03 10:19:51 +00:00
tunnel_get,
tunnel_post,
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
error_full,
rtsp-client: vmethod for modifying tunnel GET response Add a vmethod tunnel_http_response where the response to the HTTP GET for tunneled connections can be modified. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=725879
2014-03-06 12:52:02 +00:00
tunnel_lost,
tunnel_http_response
Use ASYNC RTSP io Use the async RTSP channels instead of spawning a new thread for each client. If a sessionid is specified in a request, fail if we don't have the session.
2009-02-18 17:57:31 +00:00
};
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
static void
client_watch_notify (GstRTSPClient * client)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv = client->priv;
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
gboolean closed = TRUE;
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
GST_INFO ("client %p: watch destroyed", client);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->watch = NULL;
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
/* remove all sessions if the media says so and so drop the extra client ref */
rtsp-client: Clean up watch/watch context and related state consistently And assert that it was cleaned up properly before the client is finalized. If something is still around when the client is shut down then something went very wrong before. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:16:18 +00:00
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
gst_rtsp_client_set_send_messages_func (client, NULL, NULL, NULL);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
rtsp_ctrl_timeout_remove (client);
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
gst_rtsp_client_session_filter (client, cleanup_session, &closed);
rtsp-client: Clean up watch/watch context and related state consistently And assert that it was cleaned up properly before the client is finalized. If something is still around when the client is shut down then something went very wrong before. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:16:18 +00:00
rtsp-media: Add property to decide if sending media should be stopped when a client disconnects without TEARDOWN Without TEARDOWN it might be desireable to keep the media running and continue sending data to the client, even if the RTSP connection itself is disconnected. Only do this for session medias that have only UDP transports. If there's at least on TCP transport, it will stop working and cause problems when the connection is disconnected. https://bugzilla.gnome.org/show_bug.cgi?id=758999
2015-12-09 16:24:24 +00:00
if (closed)
g_signal_emit (client, gst_rtsp_client_signals[SIGNAL_CLOSED], 0, NULL);
client: handle lost_tunnel callbacks Handle lost_tunnel callbacks and use it to store the tunnelid back into the hashtable so that we can reuse it for when the client reopens the POST socket. Close the connection after a TEARDOWN. Make sure or watchid is cleared when the watch is removed. Fixes #612915
2010-04-06 09:13:51 +00:00
g_object_unref (client);
}
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
/**
* gst_rtsp_client_attach:
* @client: a #GstRTSPClient
* @context: (allow-none): a #GMainContext
*
* Attaches @client to @context. When the mainloop for @context is run, the
rtsp-*: Refer to NULL as a constant in comments Plus one typo fix. https://bugzilla.gnome.org/show_bug.cgi?id=714988
2013-11-22 01:28:28 +00:00
* client will be dispatched. When @context is %NULL, the default context will be
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
* used).
*
* This function should be called when the client properties and urls are fully
* configured and the client is ready to start.
*
* Returns: the ID (greater than 0) for the source within the GMainContext.
*/
guint
gst_rtsp_client_attach (GstRTSPClient * client, GMainContext * context)
{
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
GstRTSPClientPrivate *priv;
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
GSource *timer_src;
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
guint res;
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
GWeakRef *client_weak_ref = g_new (GWeakRef, 1);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), 0);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv = client->priv;
docs: improve docs
2013-07-11 14:57:14 +00:00
g_return_val_if_fail (priv->connection != NULL, 0);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
g_return_val_if_fail (priv->watch == NULL, 0);
rtsp-client: Only unref client watch context on finalize, to avoid deadlock Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/issues/127 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/176>
2020-12-14 01:12:38 +00:00
g_return_val_if_fail (priv->watch_context == NULL, 0);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
client: ref the context until rtsp watch is alive Fixes https://bugzilla.gnome.org/show_bug.cgi?id=731569
2014-06-12 11:49:17 +00:00
/* make sure noone will free the context before the watch is destroyed */
priv->watch_context = g_main_context_ref (context);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
/* create watch for the connection and attach */
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
priv->watch = gst_rtsp_watch_new (priv->connection, &watch_funcs,
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
g_object_ref (client), (GDestroyNotify) client_watch_notify);
rtsp-client: Only allow to set either a send_func or send_messages_func but not both And route all messages through the send_func if no send_messages_func was provided. We otherwise break backwards compatibility.
2019-01-25 12:22:42 +00:00
gst_rtsp_client_set_send_func (client, NULL, NULL, NULL);
rtsp-client: Add support for sending buffer lists directly Fixes https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/issues/29
2018-09-17 19:18:46 +00:00
gst_rtsp_client_set_send_messages_func (client, do_send_messages, priv->watch,
(GDestroyNotify) gst_rtsp_watch_unref);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
client: make define for the WATCH_BACKLOG See https://bugzilla.gnome.org/show_bug.cgi?id=736322
2014-09-16 09:29:38 +00:00
gst_rtsp_watch_set_send_backlog (priv->watch, 0, WATCH_BACKLOG_SIZE);
rtsp-client: set the client backlog Set the client backlog to a reasonable default
2012-12-14 10:58:29 +00:00
gst-rtsp-server: fix race in rtsp-client When tunneling over HTTP, if connection on the second channel happens before the control timer is created we may trigger an assert in rtsp_ctrl_timeout_remove(). Avoid that by taking the priv->lock before attaching the client thread to the context. Fixes #1025 Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1867>
2022-03-07 08:14:46 +00:00
/* take the lock before attaching the client watch, so that the client thread
* can not access the control channel timer until it's properly in place */
g_mutex_lock (&priv->lock);
client: free watch context only once The watch context is freed when the source is destroyed. Avoids a CRITICAL when we try to unref the context twice.
2014-07-01 14:12:13 +00:00
GST_INFO ("client %p: attaching to context %p", client, context);
rtsp: make object details private Make all object details private Add methods to access private bits
2012-11-29 10:11:05 +00:00
res = gst_rtsp_watch_attach (priv->watch, context);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
/* Setting up a timeout for the RTSP control channel until a session
* is up where it is handling timeouts. */
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
/* remove old timeout if any */
rtsp_ctrl_timeout_remove_unlocked (client->priv);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
timer_src = g_timeout_source_new_seconds (RTSP_CTRL_CB_INTERVAL);
rtsp-client: Fix race condition in rtsp ctrl timeout by WeakRef client There was a race condition where client was being finalized and concurrently in some other thread the rtsp ctrl timout was relying on client data that was being freed. When rtsp ctrl timeout is setup, a WeakRef on Client is set. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/121>
2020-05-25 11:49:45 +00:00
g_weak_ref_init (client_weak_ref, client);
g_source_set_callback (timer_src, rtsp_ctrl_timeout_cb, client_weak_ref,
rtsp_ctrl_timeout_destroy_notify);
rtsp-client: Store the timeout source by pointer instead of id That way we don't have to retrieve it again from the main context when destroying it but can directly do so. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-rtsp-server/-/merge_requests/115>
2020-04-28 20:33:49 +00:00
g_source_attach (timer_src, priv->watch_context);
priv->rtsp_ctrl_timeout = timer_src;
GST_DEBUG ("rtsp control setting up session timeout %p.",
priv->rtsp_ctrl_timeout);
rtsp-client: add IDLE timeout, before session exists The RTSP server will not timeout an idle RTSP connection (note this is different from doing timeout on a RTSP session). At least for Apache this is a problem when running RTSP over HTTPS since it uses one of the threads (there is a rather limited number) that are available for handling requests. https://bugzilla.gnome.org/show_bug.cgi?id=771830
2016-11-21 15:02:39 +00:00
g_mutex_unlock (&priv->lock);
server: rework maincontext handling in clients Make a separate method to attach a client to a MainContext. Let the server decide in what GMainContext the client will operate and give this context to the client in attach. Then the server can later decide to use a separate thread for each client or just use the mainthread.
2012-11-12 13:09:09 +00:00
return res;
}
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
/**
* gst_rtsp_client_session_filter:
server: add method to iterate clients of server
2013-08-02 15:15:09 +00:00
* @client: a #GstRTSPClient
rtsp-server: sprinkle some allow-none annotations for g-i
2013-11-18 10:47:04 +00:00
* @func: (scope call) (allow-none): a callback
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
* @user_data: user data passed to @func
*
* Call @func for each session managed by @client. The result value of @func
* determines what happens to the session. @func will be called with @client
* locked so no further actions on @client can be performed from @func.
*
* If @func returns #GST_RTSP_FILTER_REMOVE, the session will be removed from
* @client.
*
* If @func returns #GST_RTSP_FILTER_KEEP, the session will remain in @client.
*
* If @func returns #GST_RTSP_FILTER_REF, the session will remain in @client but
* will also be added with an additional ref to the result #GList of this
* function..
*
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
* When @func is %NULL, #GST_RTSP_FILTER_REF will be assumed for each session.
*
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
* Returns: (element-type GstRTSPSession) (transfer full): a #GList with all
* sessions for which @func returned #GST_RTSP_FILTER_REF. After usage, each
* element in the #GList should be unreffed before the list is freed.
*/
GList *
gst_rtsp_client_session_filter (GstRTSPClient * client,
GstRTSPClientSessionFilterFunc func, gpointer user_data)
{
GstRTSPClientPrivate *priv;
GList *result, *walk, *next;
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
GHashTable *visited;
guint cookie;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
g_return_val_if_fail (GST_IS_RTSP_CLIENT (client), NULL);
priv = client->priv;
result = NULL;
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
if (func)
visited = g_hash_table_new_full (NULL, NULL, g_object_unref, NULL);
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
g_mutex_lock (&priv->lock);
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
restart:
cookie = priv->sessions_cookie;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
for (walk = priv->sessions; walk; walk = next) {
GstRTSPSession *sess = walk->data;
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
GstRTSPFilterResult res;
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
gboolean changed;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
next = g_list_next (walk);
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
if (func) {
/* only visit each session once */
if (g_hash_table_contains (visited, sess))
continue;
g_hash_table_add (visited, g_object_ref (sess));
g_mutex_unlock (&priv->lock);
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
res = func (client, sess, user_data);
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
g_mutex_lock (&priv->lock);
} else
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
res = GST_RTSP_FILTER_REF;
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
changed = (cookie != priv->sessions_cookie);
rtsp: allow NULL func in filters Passing a null function make the filters return a list of refcounted objects.
2013-11-15 15:35:05 +00:00
switch (res) {
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
case GST_RTSP_FILTER_REMOVE:
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
/* stop watching the session and pretend it went away, if the list was
* changed, we can't use the current list position, try to see if we
* still have the session */
client_unwatch_session (client, sess, changed ? NULL : walk);
cookie = priv->sessions_cookie;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
break;
case GST_RTSP_FILTER_REF:
result = g_list_prepend (result, g_object_ref (sess));
break;
case GST_RTSP_FILTER_KEEP:
default:
break;
}
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
if (changed)
goto restart;
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
}
g_mutex_unlock (&priv->lock);
filter: Release lock in filter functions Release the object lock before calling the filter functions. We need to keep a cookie to detect when the list changed during the filter callback. We also keep a hashtable to make sure we only call the filter function once for each object in case of concurrent modification. Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732950
2014-07-10 09:32:20 +00:00
if (func)
g_hash_table_unref (visited);
client: add method to filter managed sessions Add a method to filter the sessions managed by this client connection. See https://bugzilla.gnome.org/show_bug.cgi?id=703016
2013-06-26 15:18:33 +00:00
return result;
}
Reference in a new issue Copy permalink