asterisk/res/ari/ari_websockets.c

202 lines
4.8 KiB
C
Raw Normal View History

/*
* Asterisk -- An open source telephony toolkit.
*
* Copyright (C) 2013, Digium, Inc.
*
* David M. Lee, II <dlee@digium.com>
*
* See http://www.asterisk.org for more information about
* the Asterisk project. Please do not directly contact
* any of the maintainers of this project for assistance;
* the project provides a web site, mailing lists and IRC
* channels for your use.
*
* This program is free software, distributed under the terms of
* the GNU General Public License Version 2. See the LICENSE file
* at the top of the source tree.
*/
#include "asterisk.h"
#include "asterisk/ari.h"
optional_api: Fix linking problems between modules that export global symbols With the new work in Asterisk 12, there are some uses of the optional_api that are prone to failure. The details are rather involved, and captured on [the wiki][1]. This patch addresses the issue by removing almost all of the magic from the optional API implementation. Instead of relying on weak symbol resolution, a new optional_api.c module was added to Asterisk core. For modules providing an optional API, the pointer to the implementation function is registered with the core. For modules that use an optional API, a pointer to a stub function, along with a optional_ref function pointer are registered with the core. The optional_ref function pointers is set to the implementation function when it's provided, or the stub function when it's now. Since the implementation no longer relies on magic, it is now supported on all platforms. In the spirit of choice, an OPTIONAL_API flag was added, so we can disable the optional_api if needed (maybe it's buggy on some bizarre platform I haven't tested on) The AST_OPTIONAL_API*() macros themselves remained unchanged, so existing code could remain unchanged. But to help with debugging the optional_api, the patch limits the #include of optional API's to just the modules using the API. This also reduces resource waste maintaining optional_ref pointers that aren't used. Other changes made as a part of this patch: * The stubs for http_websocket that wrap system calls set errno to ENOSYS. * res_http_websocket now properly increments module use count. * In loader.c, the while() wrappers around dlclose() were removed. The while(!dlclose()) is actually an anti-pattern, which can lead to infinite loops if the module you're attempting to unload exports a symbol that was directly linked to. * The special handling of nonoptreq on systems without weak symbol support was removed, since we no longer rely on weak symbols for optional_api. [1]: https://wiki.asterisk.org/wiki/x/wACUAQ (closes issue ASTERISK-22296) Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/2797/ ........ Merged revisions 397989 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397990 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-08-30 13:40:27 +00:00
#include "asterisk/astobj2.h"
#include "asterisk/http_websocket.h"
#include "internal.h"
/*! \file
*
* \brief WebSocket support for RESTful API's.
* \author David M. Lee, II <dlee@digium.com>
*/
struct ast_ari_websocket_session {
struct ast_websocket *ws_session;
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
int (*validator)(struct ast_json *);
};
static void websocket_session_dtor(void *obj)
{
struct ast_ari_websocket_session *session = obj;
ast_websocket_unref(session->ws_session);
session->ws_session = NULL;
}
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
/*!
* \brief Validator that always succeeds.
*/
static int null_validator(struct ast_json *json)
{
return 1;
}
struct ast_ari_websocket_session *ast_ari_websocket_session_create(
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
struct ast_websocket *ws_session, int (*validator)(struct ast_json *))
{
RAII_VAR(struct ast_ari_websocket_session *, session, NULL, ao2_cleanup);
res_http_websocket: Close websocket correctly and use careful fwrite When a client takes a long time to process information received from Asterisk, a write operation using fwrite may fail to write all information. This causes the underlying file stream to be in an unknown state, such that the socket must be disconnected. Unfortunately, there are two problems with this in Asterisk's existing websocket code: 1. Periodically, during the read loop, Asterisk must write to the connected websocket to respond to pings. As such, Asterisk maintains a reference to the session during the loop. When ast_http_websocket_write fails, it may cause the session to decrement its ref count, but this in and of itself does not break the read loop. The read loop's write, on the other hand, does not break the loop if it fails. This causes the socket to get in a 'stuck' state, preventing the client from reconnecting to the server. 2. More importantly, however, is that the fwrite in ast_http_websocket_write fails with a large volume of data when the client takes awhile to process the information. When it does fail, it fails writing only a portion of the bytes. With some debugging, it was shown that this was failing in a similar fashion to ASTERISK-12767. Switching this over to ast_careful_fwrite with a long enough timeout solved the problem. Note that this version of the patch, unlike r417310 in Asterisk 11, exposes configuration options beyond just chan_sip's sip.conf. Configuration options to configure the write timeout have also been added to pjsip.conf and ari.conf. #ASTERISK-23917 #close Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/3624/ ........ Merged revisions 417310 from http://svn.asterisk.org/svn/asterisk/branches/11 ........ Merged revisions 417311 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@417317 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2014-06-26 12:21:14 +00:00
RAII_VAR(struct ast_ari_conf *, config, ast_ari_config_get(), ao2_cleanup);
if (ws_session == NULL) {
return NULL;
}
res_http_websocket: Close websocket correctly and use careful fwrite When a client takes a long time to process information received from Asterisk, a write operation using fwrite may fail to write all information. This causes the underlying file stream to be in an unknown state, such that the socket must be disconnected. Unfortunately, there are two problems with this in Asterisk's existing websocket code: 1. Periodically, during the read loop, Asterisk must write to the connected websocket to respond to pings. As such, Asterisk maintains a reference to the session during the loop. When ast_http_websocket_write fails, it may cause the session to decrement its ref count, but this in and of itself does not break the read loop. The read loop's write, on the other hand, does not break the loop if it fails. This causes the socket to get in a 'stuck' state, preventing the client from reconnecting to the server. 2. More importantly, however, is that the fwrite in ast_http_websocket_write fails with a large volume of data when the client takes awhile to process the information. When it does fail, it fails writing only a portion of the bytes. With some debugging, it was shown that this was failing in a similar fashion to ASTERISK-12767. Switching this over to ast_careful_fwrite with a long enough timeout solved the problem. Note that this version of the patch, unlike r417310 in Asterisk 11, exposes configuration options beyond just chan_sip's sip.conf. Configuration options to configure the write timeout have also been added to pjsip.conf and ari.conf. #ASTERISK-23917 #close Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/3624/ ........ Merged revisions 417310 from http://svn.asterisk.org/svn/asterisk/branches/11 ........ Merged revisions 417311 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@417317 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2014-06-26 12:21:14 +00:00
if (config == NULL || config->general == NULL) {
return NULL;
}
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
if (validator == NULL) {
validator = null_validator;
}
if (ast_websocket_set_nonblock(ws_session) != 0) {
ast_log(LOG_ERROR,
"ARI web socket failed to set nonblock; closing: %s\n",
strerror(errno));
return NULL;
}
res_http_websocket: Close websocket correctly and use careful fwrite When a client takes a long time to process information received from Asterisk, a write operation using fwrite may fail to write all information. This causes the underlying file stream to be in an unknown state, such that the socket must be disconnected. Unfortunately, there are two problems with this in Asterisk's existing websocket code: 1. Periodically, during the read loop, Asterisk must write to the connected websocket to respond to pings. As such, Asterisk maintains a reference to the session during the loop. When ast_http_websocket_write fails, it may cause the session to decrement its ref count, but this in and of itself does not break the read loop. The read loop's write, on the other hand, does not break the loop if it fails. This causes the socket to get in a 'stuck' state, preventing the client from reconnecting to the server. 2. More importantly, however, is that the fwrite in ast_http_websocket_write fails with a large volume of data when the client takes awhile to process the information. When it does fail, it fails writing only a portion of the bytes. With some debugging, it was shown that this was failing in a similar fashion to ASTERISK-12767. Switching this over to ast_careful_fwrite with a long enough timeout solved the problem. Note that this version of the patch, unlike r417310 in Asterisk 11, exposes configuration options beyond just chan_sip's sip.conf. Configuration options to configure the write timeout have also been added to pjsip.conf and ari.conf. #ASTERISK-23917 #close Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/3624/ ........ Merged revisions 417310 from http://svn.asterisk.org/svn/asterisk/branches/11 ........ Merged revisions 417311 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@417317 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2014-06-26 12:21:14 +00:00
if (ast_websocket_set_timeout(ws_session, config->general->write_timeout)) {
ast_log(LOG_WARNING, "Failed to set write timeout %d on ARI web socket\n",
config->general->write_timeout);
}
session = ao2_alloc(sizeof(*session), websocket_session_dtor);
if (!session) {
return NULL;
}
ao2_ref(ws_session, +1);
session->ws_session = ws_session;
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
session->validator = validator;
ao2_ref(session, +1);
return session;
}
struct ast_json *ast_ari_websocket_session_read(
struct ast_ari_websocket_session *session)
{
RAII_VAR(struct ast_json *, message, NULL, ast_json_unref);
if (ast_websocket_fd(session->ws_session) < 0) {
return NULL;
}
while (!message) {
int res;
char *payload;
uint64_t payload_len;
enum ast_websocket_opcode opcode;
int fragmented;
res = ast_wait_for_input(
ast_websocket_fd(session->ws_session), -1);
if (res <= 0) {
ast_log(LOG_WARNING, "WebSocket poll error: %s\n",
strerror(errno));
return NULL;
}
res = ast_websocket_read(session->ws_session, &payload,
&payload_len, &opcode, &fragmented);
if (res != 0) {
ast_log(LOG_WARNING, "WebSocket read error: %s\n",
strerror(errno));
return NULL;
}
switch (opcode) {
case AST_WEBSOCKET_OPCODE_CLOSE:
ast_debug(1, "WebSocket closed\n");
return NULL;
case AST_WEBSOCKET_OPCODE_TEXT:
message = ast_json_load_buf(payload, payload_len, NULL);
if (message == NULL) {
ast_log(LOG_WARNING,
"WebSocket input failed to parse\n");
}
break;
default:
/* Ignore all other message types */
break;
}
}
return ast_json_ref(message);
}
#define VALIDATION_FAILED \
"{" \
" \"error\": \"InvalidMessage\"," \
" \"message\": \"Message validation failed\"" \
ARI: Channels added to Stasis application during WebSocket creation ... Prior to ASTERISK-24988, the WebSocket handshake was resolved before Stasis applications were registered. This was done such that the WebSocket would be ready when an application is registered. However, by creating the WebSocket first, the client had the ability to make requests for the Stasis application it thought had been created with the initial handshake request. The inevitable conclusion of this scenario was the cart being put before the horse. ASTERISK-24988 resolved half of the problem by ensuring that the applications were created and registered with Stasis prior to completing the handshake with the client. While this meant that Stasis was ready when the client received the green-light from Asterisk, it also meant that the WebSocket was not yet ready for Stasis to dispatch messages. This patch introduces a message queuing mechanism for delaying messages from Stasis applications while the WebSocket is being constructed. When the ARI event processor receives the message from the WebSocket that it is being created, the event processor instantiates an event session which contains a message queue. It then tries to create and register the requested applications with Stasis. Messages that are dispatched from Stasis between this point and the point at which the event processor is notified the WebSocket is ready, are stashed in the queue. Once the WebSocket has been built, the queue's messages are dispatched in the order in which they were originally received and the queue is concurrently cleared. ASTERISK-25181 #close Reported By: Matt Jordan Change-Id: Iafef7b85a2e0bf78c114db4c87ffc3d16d671a17
2015-07-31 16:27:23 +00:00
"}"
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
int ast_ari_websocket_session_write(struct ast_ari_websocket_session *session,
struct ast_json *message)
{
RAII_VAR(char *, str, NULL, ast_json_free);
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
#ifdef AST_DEVMODE
if (!session->validator(message)) {
ast_log(LOG_ERROR, "Outgoing message failed validation\n");
res_http_websocket: Avoid passing strlen() to ast_websocket_write(). We have seen a rash of test failures on a 32-bit build agent. Commit 48698a5e21d7307f61b5fb2bd39fd593bc1423ca solved an obvious problem where we were not encoding a 64-bit value correctly over the wire. This commit, however, did not solve the test failures. In the failing tests, ARI is attempting to send a 537 byte text frame over a websocket. When sending a frame this small, 16 bits are all that is required in order to encode the payload length on the websocket frame. However, ast_websocket_write() thinks that the payload length is greater than 65535 and therefore writes out a 64 bit payload length. Inspecting this payload length, the lower 32 bits are exactly what we would expect it to be, 537 in hex. The upper 32 bits, are junk values that are not expected to be there. In the failure, we are passing the result of strlen() to a function that expects a uint64_t parameter to be passed in. strlen() returns a size_t, which on this 32-bit machine is 32 bits wide. Normally, passing a 32-bit unsigned value to somewhere where a 64-bit unsigned value is expected would cause no problems. In fact, in manual runs of failing tests, this works just fine. However, ast_websocket_write() uses the Asterisk optional API, which means that rather than a simple function call, there are a series of macros that are used for its declaration and implementation. These macros may be causing some sort of error to occur when converting from a 32 bit quantity to a 64 bit quantity. This commit changes the logic by making existing ast_websocket_write() calls use ast_websocket_write_string() instead. Within ast_websocket_write_string(), the 64-bit converted strlen is saved in a local variable, and that variable is passed to ast_websocket_write() instead. Note that this commit message is full of speculation rather than certainty. This is because the observed test failures, while always present in automated test runs, never occur when tests are manually attempted on the same test agent. The idea behind this commit is to fix a theoretical issue by performing changes that should, at the least, cause no harm. If it turns out that this change does not fix the failing tests, then this commit should be reverted. Change-Id: I4458dd87d785ca322b89c152b223a540a3d23e67
2015-08-03 16:06:07 +00:00
return ast_websocket_write_string(session->ws_session, VALIDATION_FAILED);
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
}
#endif
str = ast_json_dump_string_format(message, ast_ari_json_format());
if (str == NULL) {
ast_log(LOG_ERROR, "Failed to encode JSON object\n");
return -1;
}
#ifdef AST_DEVMODE
ast_debug(3, "Examining ARI event (length %u): \n%s\n", (unsigned int) strlen(str), str);
#endif
res_http_websocket: Avoid passing strlen() to ast_websocket_write(). We have seen a rash of test failures on a 32-bit build agent. Commit 48698a5e21d7307f61b5fb2bd39fd593bc1423ca solved an obvious problem where we were not encoding a 64-bit value correctly over the wire. This commit, however, did not solve the test failures. In the failing tests, ARI is attempting to send a 537 byte text frame over a websocket. When sending a frame this small, 16 bits are all that is required in order to encode the payload length on the websocket frame. However, ast_websocket_write() thinks that the payload length is greater than 65535 and therefore writes out a 64 bit payload length. Inspecting this payload length, the lower 32 bits are exactly what we would expect it to be, 537 in hex. The upper 32 bits, are junk values that are not expected to be there. In the failure, we are passing the result of strlen() to a function that expects a uint64_t parameter to be passed in. strlen() returns a size_t, which on this 32-bit machine is 32 bits wide. Normally, passing a 32-bit unsigned value to somewhere where a 64-bit unsigned value is expected would cause no problems. In fact, in manual runs of failing tests, this works just fine. However, ast_websocket_write() uses the Asterisk optional API, which means that rather than a simple function call, there are a series of macros that are used for its declaration and implementation. These macros may be causing some sort of error to occur when converting from a 32 bit quantity to a 64 bit quantity. This commit changes the logic by making existing ast_websocket_write() calls use ast_websocket_write_string() instead. Within ast_websocket_write_string(), the 64-bit converted strlen is saved in a local variable, and that variable is passed to ast_websocket_write() instead. Note that this commit message is full of speculation rather than certainty. This is because the observed test failures, while always present in automated test runs, never occur when tests are manually attempted on the same test agent. The idea behind this commit is to fix a theoretical issue by performing changes that should, at the least, cause no harm. If it turns out that this change does not fix the failing tests, then this commit should be reverted. Change-Id: I4458dd87d785ca322b89c152b223a540a3d23e67
2015-08-03 16:06:07 +00:00
if (ast_websocket_write_string(session->ws_session, str)) {
ast_log(LOG_NOTICE, "Problem occurred during websocket write, websocket closed\n");
return -1;
}
return 0;
}
optional_api: Fix linking problems between modules that export global symbols With the new work in Asterisk 12, there are some uses of the optional_api that are prone to failure. The details are rather involved, and captured on [the wiki][1]. This patch addresses the issue by removing almost all of the magic from the optional API implementation. Instead of relying on weak symbol resolution, a new optional_api.c module was added to Asterisk core. For modules providing an optional API, the pointer to the implementation function is registered with the core. For modules that use an optional API, a pointer to a stub function, along with a optional_ref function pointer are registered with the core. The optional_ref function pointers is set to the implementation function when it's provided, or the stub function when it's now. Since the implementation no longer relies on magic, it is now supported on all platforms. In the spirit of choice, an OPTIONAL_API flag was added, so we can disable the optional_api if needed (maybe it's buggy on some bizarre platform I haven't tested on) The AST_OPTIONAL_API*() macros themselves remained unchanged, so existing code could remain unchanged. But to help with debugging the optional_api, the patch limits the #include of optional API's to just the modules using the API. This also reduces resource waste maintaining optional_ref pointers that aren't used. Other changes made as a part of this patch: * The stubs for http_websocket that wrap system calls set errno to ENOSYS. * res_http_websocket now properly increments module use count. * In loader.c, the while() wrappers around dlclose() were removed. The while(!dlclose()) is actually an anti-pattern, which can lead to infinite loops if the module you're attempting to unload exports a symbol that was directly linked to. * The special handling of nonoptreq on systems without weak symbol support was removed, since we no longer rely on weak symbols for optional_api. [1]: https://wiki.asterisk.org/wiki/x/wACUAQ (closes issue ASTERISK-22296) Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/2797/ ........ Merged revisions 397989 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397990 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-08-30 13:40:27 +00:00
void ari_handle_websocket(struct ast_websocket_server *ws_server,
struct ast_tcptls_session_instance *ser, const char *uri,
enum ast_http_method method, struct ast_variable *get_params,
struct ast_variable *headers)
{
struct ast_http_uri fake_urih = {
.data = ws_server,
};
ast_websocket_uri_cb(ser, &fake_urih, uri, method, get_params,
headers);
}
ARI: Channels added to Stasis application during WebSocket creation ... Prior to ASTERISK-24988, the WebSocket handshake was resolved before Stasis applications were registered. This was done such that the WebSocket would be ready when an application is registered. However, by creating the WebSocket first, the client had the ability to make requests for the Stasis application it thought had been created with the initial handshake request. The inevitable conclusion of this scenario was the cart being put before the horse. ASTERISK-24988 resolved half of the problem by ensuring that the applications were created and registered with Stasis prior to completing the handshake with the client. While this meant that Stasis was ready when the client received the green-light from Asterisk, it also meant that the WebSocket was not yet ready for Stasis to dispatch messages. This patch introduces a message queuing mechanism for delaying messages from Stasis applications while the WebSocket is being constructed. When the ARI event processor receives the message from the WebSocket that it is being created, the event processor instantiates an event session which contains a message queue. It then tries to create and register the requested applications with Stasis. Messages that are dispatched from Stasis between this point and the point at which the event processor is notified the WebSocket is ready, are stashed in the queue. Once the WebSocket has been built, the queue's messages are dispatched in the order in which they were originally received and the queue is concurrently cleared. ASTERISK-25181 #close Reported By: Matt Jordan Change-Id: Iafef7b85a2e0bf78c114db4c87ffc3d16d671a17
2015-07-31 16:27:23 +00:00
const char *ast_ari_websocket_session_id(
const struct ast_ari_websocket_session *session)
{
return ast_websocket_session_id(session->ws_session);
}