There are some situations in Asterisk where ast_frame and/or iax_frame
structures are rapidly allocatted and freed (at least 50 times per second
for one call).
This code significantly improves the performance of ast_frame_header_new(),
ast_frdup(), ast_frfree(), iax_frame_new(), and iax_frame_free() by keeping
a thread-local cache of these structures and using frames from the cache
whenever possible instead of calling malloc/free every time.
This commit also converts the ast_frame and iax_frame structures to use the
linked list macros.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@41278 65c4cc65-6c06-0410-ace0-fbb531ad65f3
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r40994 | russell | 2006-08-24 15:41:26 -0400 (Thu, 24 Aug 2006) | 11 lines
Fix a few issues related to the handling of channel variables
- in pbx_builtin_serialize_variables(), the variable list traversal would stop
on a variables with empty name/values, which is not appropriate
- When removing the GROUP variables, use AST_LIST_REMOVE_CURRENT instead of
AST_LIST_REMOVE
- During masquerading, when copying the variables list from one channel to the
other, using AST_LIST_INSERT_TAIL is not valid for appending a whole list.
It leaves the tail pointer of the list invalid. Introduce a new macro,
AST_LIST_APPEND_LIST that appends a list properly.
(issue #7802, softins)
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@40995 65c4cc65-6c06-0410-ace0-fbb531ad65f3
also, keep trying to dlclose() a module until it actually goes away, since it may have other modules it brought in when it was loaded (thanks PCadach for pointing this problem out to me)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@40949 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- add ast_asprintf() and ast_vasprintf()
- tweak doxygen comments
- simplify the definition of a flag macro
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@40775 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- restructured build tree and makefiles to eliminate recursion problems
- support for embedded modules
- support for static builds
- simpler cross-compilation support
- simpler module/loader interface (no exported symbols)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@40722 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- instead of defining a free() wrapper in a bunch of files, define it as
ast_free() in utils.h and remove the copies from all the files.
- centralize and abstract the code used for doing thread storage. The code
lives in threadstorage.h, with one function being implemented in utils.c.
This new API includes generic thread storage as well as special functions
for handling thread local dynamic length string buffers.
- update ast_inet_ntoa() to use the new threadstorage API
- update ast_state2str() to use the new threadstorage API
- update ast_cli() to use the new threadstorage API
- Modify manager_event() to use thread storage. Instead of using a buffer of
4096 characters as the workspace for building the manager event, use a thread
local dynamic string. Now there is no length limitation on the length of the
body of a manager event.
- Significantly simplify the handling of ast_verbose() ...
- Instead of using a static char buffer and a lock to make sure only one
thread can be using ast_verbose() at a time, use a thread local dynamic
string as the workspace for preparing the verbose message. Instead of
locking around the entire function, the only locking done now is when the
message has been built and is being deliviered to the list of registered
verbose message handlers.
- This function was doing a strdup() on every message passed to it and
keeping a queue of the last 200 messages in memory. This has been
completely removed. The only place this was used was that if there were
any messages in the verbose queue when a verbose handler was registered,
all of the messages in the queue would be fed to it. So, I just made sure
that the console verbose handler and the network verbose handler (for
remote asterisk consoles) were registered before any verbose messages.
pbx_gtkconsole and pbx_kdeconsole will now lose a few verbose messages at
startup, but I didn't feel the performance hit of this message queue was
worth saving the initial verbose output for these very rarely used modules.
- I have removed the last three arguments to the verbose handlers, leaving
only the string itself because they aren't needed anymore. For example,
ast_verbose had some logic for telling the verbose handler to add
a newline if the buffer was completely full. Now that the buffer can grow
as needed, this doesn't matter anymore.
- remove unused function, ast_verbose_dmesg() which was to dispatch the
message queue
- Convert the list of verbose handlers to use the linked list macros.
- add missing newline characters to a few ast_verbose() calls
- convert the list of log channels to use the linked list macros in logger.c
- fix close_logger() to close all of the files it opened for logging
- update ast_log() to use a thread local dynamic string for its workspace
for preparing log messages instead of a buffer of size BUFSIZ (8kB on my
system) allocated on the stack. The dynamic string in this case is limited
to only growing to a maximum size of BUFSIZ.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@39272 65c4cc65-6c06-0410-ace0-fbb531ad65f3
int to string or string to int operations.
"pure" essentially says that this function has no side effects aside from its
result, and the result depends on nothing else other than its arguments and
global variables. "const" is a more strict form of "pure", where the function
also doesn't access any global variables.
From the gcc manual: "Such a function can be subject to common subexpression
elimination and loop optimization just as an arithmetic operator would be."
This also tells the compiler that it is safe to call the function fewer times
than the code says to, given the same arguments, since the result will always
be the same.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38452 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- Fix some breakage I introduced a while ago that made the timestamps option
not functional for CLI verbose output.
- Remove the use of the timestamps option for log output, since it was not
functional.
- clarify text referring to the timestamps option so that it is clear that it
only applies to CLI verbose output
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38439 65c4cc65-6c06-0410-ace0-fbb531ad65f3
make read/write/hold work on samples, not bytes
add an API call to find out how many samples are available in a slinfactory
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38418 65c4cc65-6c06-0410-ace0-fbb531ad65f3
fix prototype for a channel walking function to use a const input pointer
use existing channel walk by name prefix instead of reproducing that code in this app
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38389 65c4cc65-6c06-0410-ace0-fbb531ad65f3
if the channel is already in the autoservice list.
Why is this a valid case to return -1, you ask? Well, there should never be
any code where it is not clear if the channel is in autoservice or not because
trying to read frames from a channel that is in the autoservice list will lead
to bad results because more than one thread will be waiting on frames to arrive
on the channel and then trying to read them.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38076 65c4cc65-6c06-0410-ace0-fbb531ad65f3
inet_ntoa, which uses thread specific data (aka thread local storage) instead
of stack allocatted buffers to store the result.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@38042 65c4cc65-6c06-0410-ace0-fbb531ad65f3
split support for G726-32 into RFC3551 and AAL2 packing orders, since both are in use
change "G726-32" to be RFC3551 packing order, in spite of devices that use AAL2 order with this MIME type
add ability to directly transcode between packing orders
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@37494 65c4cc65-6c06-0410-ace0-fbb531ad65f3
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
support the new location for zaptel.h and tonezone.h
use the dependency information output by menuselect to build Makefile rules for each module for header files and libraries
combine the common rules into a top-level Makefile.rules file
remove all (now) unnecessary stuff from subdir Makefiles
change translator API so that the newpvt() callback returns an int instead of a pointer (it no longer allocates memory)
alphabetize --with-<foo> options in configure script
enhance Net-SNMP support in configure script to provide a --with-netsnmp option
fix support for --with-pq so that if pg-config is not found when --with-pq is specified, an error will be generated
add 'optional package' usage to modules now that menuselect can output it
allow res_snmp to build by default, since the new loader changes coming soon will solve the function naming problem (and users can disable it via menuselect anyway)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@35832 65c4cc65-6c06-0410-ace0-fbb531ad65f3
allocated. These changes caused crashes when using a channel type that did
not support the jitterbuffer. Instead of fixing why it's crashing, I'm going
to implement this in a better way next week. The way I did it caused a
jitterbuffer to be allocated on every channel where the channel type supported
jitterbuffers, even if they were disabled.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@35746 65c4cc65-6c06-0410-ace0-fbb531ad65f3
so that channels not using a jitterbuffer don't waste as much memory
- ensure that the channel drivers that use jitterbuffers can handle a failure
from configuring a jitterbuffer on a new channel because of a memory
allocation error
- On passing through these channel drivers, configure the jitterbuffer before
starting the PBX thread instead of afterwards. If the pbx fails to start for
whatever reason, this would have caused a crash.
- Also on passing, move the increase of the usecount to after all of the
possible failure conditions in the function
- fix a place where ast_update_use_count() was not called
- ensure that the owner channel pointer of the channel pvt strcutures is set to
NULL in failure conditions
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@35553 65c4cc65-6c06-0410-ace0-fbb531ad65f3
and some patches (all disclaimed).
- Don't change RTP properties if we reject a re-INVITE
- Don't add video to an outbound channel if there's no video on the inbound channel
- Don't include video in the "preferred codec" list for codec selection
- Clean up and document code that parses and adds SDP attachments
Since we do not transcode video, we can't handle video the same way as audio. This is a
bug fix patch. In future releases, we need to work on a solution for video negotiation,
not codecs but formats and framerates instead.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@32597 65c4cc65-6c06-0410-ace0-fbb531ad65f3
a new implementation of a fixed size jitterbuffer, as well as support for the
existing adaptive jitterbuffer implementation. (issue #3854, Slav Klenov)
Thank you very much to Slav Klenov of Securax and all of the people involved
in the testing of this feature for all of your hard work!
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@31052 65c4cc65-6c06-0410-ace0-fbb531ad65f3
output to remove consoles. The prototypes added to logger.h still need
doxygen documentation, as well.
- Add the new command line option to the man page
- make the mute option a flag instead of an int since it is only a binary
option
- remove useless extern keywords for prototypes added to logger.h
- rename ast_console_mute() to ast_console_toggle_mute() since that is what
it actually does
- actually apply the mute option to newly created remote consoles instead of
only working when the CLI command is used
- don't imply the NO_FORK option if the mute command line option is provided
- place the new CLI command in the correct place in the list which has to be
in alphabetical order
- Finally, clean up a few spacing issues to conform to the coding guidelines
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@30630 65c4cc65-6c06-0410-ace0-fbb531ad65f3
is an error executing the AGI script, or the AGI script itself returns a
non-zero value, the AGISTATUS variable will now be set to FAILURE instead of
SUCCESS.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@30328 65c4cc65-6c06-0410-ace0-fbb531ad65f3
update ast_mutex_init to allow mutexes that are all zero bytes to be initialized (in the case of a dynamically-allocated structure containing a mutex)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@29727 65c4cc65-6c06-0410-ace0-fbb531ad65f3
allow native bridging of RTP sessions that are not carrying DTMF even when the bridge needs to listen to DTMF (when SIP INFO is used for DTMF, for example)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@27559 65c4cc65-6c06-0410-ace0-fbb531ad65f3
So, I have removed all of the uses of AST_LIST_HEAD_INIT and replaced them
with the equivalent static initializations.
- On passing, fix a memory leak in the unload_module() function of chan_agent.
The agents list mutex was never destroyed, and the elements in the agents
list were not freed.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@26990 65c4cc65-6c06-0410-ace0-fbb531ad65f3
instead of being added to the compiler commands. This header file will be
installed and modules built outside of the main tree will be able to use the
same build options used to build the rest of Asterisk.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@26808 65c4cc65-6c06-0410-ace0-fbb531ad65f3
update iax2_indicate to pass control frame payload to the connected channel
add an API call for sending an indication with payload, and use it for control frames with payload
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@26417 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- remove some checks of the result of ast_mutex_lock, since it is not necessary
(this would be a good project to add to the janitor projects list).
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@25443 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- we can't use ast_true here because non-empty strings would no longer be
evaluated as true
document the return values of pbx_checkcondition() in doxygen format
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@25411 65c4cc65-6c06-0410-ace0-fbb531ad65f3
on the frame counters. Document it in the header file.
- provide a single exit point for a function;
- mark XXX some unclear parts of the code.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@21933 65c4cc65-6c06-0410-ace0-fbb531ad65f3
equivalent (the reason is, when passing these strings through a
statically allocated buffer, we have no way to tell between NULL and ""
so we would be unable to preserve the difference, if any).
No code changes yet.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@21743 65c4cc65-6c06-0410-ace0-fbb531ad65f3
wrappers around the basic 'say' functions, and redeclare these
wrappers as ordinary functions rather than function pointers.
This way, alternative implementations of the 'say' functions
will only have to implement the basic functions and not the
wrappers.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@21338 65c4cc65-6c06-0410-ace0-fbb531ad65f3
don't transcode via SLINEAR when the option is enabled but there is a direct path from the source to the destination
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@20962 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- Update lock.h with definitions of ast_channel_lock, ast_channel_unlock and ast_channel_trylock
- Convert some functions (but not all) in channel.c
- Fix some bugs in chan_sip.c
- Convert rest of chan_sip.c
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@20295 65c4cc65-6c06-0410-ace0-fbb531ad65f3
when you have channel locking issues.
(Part of the SIP transfer patch, where I had a *lot* of
channel locking problems)
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@20264 65c4cc65-6c06-0410-ace0-fbb531ad65f3
As partly documented in loader.c and include/asterisk/module.h,
modules are now expected to return all of their methods and flags
into a structure 'mod_data', and are normally loaded with RTLD_NOW
| RTLD_LOCAL, so symbols are resolved immediately and conflicts
should be less likely. Only in a small number of cases (res_*,
typically) modules are loaded RTLD_GLOBAL, so they can export
symbols.
The core of the change is only the two files loader.c and
include/asterisk/module.h, all the rest is simply adaptation of the
existing modules to the new API, a rather mechanical (but believe
me, time and finger-consuming!) process whose detail you can figure
out by svn diff'ing any single module.
Expect some minor compilation issue after this change, please
report it on mantis http://bugs.digium.com/view.php?id=6968
so we collect all the feedback in one place.
I am just sorry that this change missed SVN version number 20000!
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@20003 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This fixes the compilation on OS/X (the change exposed a wrong
assumption on mutex types on OS/X), but still leaves open the
bugs in initializing mutex on bsd systems, which you will see
reported as 'locking failures' on certain operations.
I need to investigate the issue further, but the best thing
i can do now is leave things as they have been for months.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@19973 65c4cc65-6c06-0410-ace0-fbb531ad65f3
These are momstly debugging tools for developers,
a bit documented in the header files (utils.h),
although more documentation is definitely necessary.
The performance impact is close to zero(*) so there is no
need to compile it conditionally.
(*) not completely true - thread destruction still needs
to search a list _but_ this can be easily optimized if we
end up with hundreds of active threads (in which case, though,
the problem is clearly elsewhere).
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@19544 65c4cc65-6c06-0410-ace0-fbb531ad65f3
- misspelled ast_mutex_logger() instead of __ast_mutex_logger()
- misplaced #define ast_mutex_init(pmutex)
- wrong arguments to __ast_mutex_logger() in one instance.
Clearly this code is too spaghetti!
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@19396 65c4cc65-6c06-0410-ace0-fbb531ad65f3
ast_register_atexit()/ ast_unregister_atexit() into asterisk.h
These are general functions, not restricted to modules, so move
them in a more proper place.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@19223 65c4cc65-6c06-0410-ace0-fbb531ad65f3
new-style modules using static symbols.
Everything will still work as before, but new-style modules
can now be defined by putting a '#define STATIC_MODULE' somewhere
before including module.h, then declaring STATIC_MODULE the
various methods (load, unload, key...) that the module is
supposed to supply, and adding a 'STD_MOD(MOD_1, reload_fn, NULL, NULL)'
macro call at the end.
A module compiled in this way will be loaded RTLD_NOW|RTLD_LOCAL
so symbol pollution is reduced, and symbols are resolved immediately.
Removing just the '#define STATIC_MODULE' will restore the old
behaviour.
In order for a module to be loaded RTLD_NOW|RTLD_LOCAL, it must not
export any symbol[1], and all the modules it depends on (e.g. res_*)
must be loaded already.
[1] Mechanisms are in place, and will be enabled later, to still
allow such modules to 'export' symbols and resolving the dependencies
irrespective of the load order.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@17790 65c4cc65-6c06-0410-ace0-fbb531ad65f3
collecting common functions in a single place and removing
them from the individual handlers.
The full description is on mantis,
http://bugs.digium.com/view.php?id=6375
and only the ogg_vorbis handler needs to be converted to
the new structure.
As a result of this change, format_au.c and format_pcm_alaw.c
should go away (in a separate commit) as their functionality
(trivial) has been merged in another file.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@17243 65c4cc65-6c06-0410-ace0-fbb531ad65f3
in pbx_exec is always 1 so it can be removed.
This change also takes away ast_exec_extension(), and lets all
switch functions (exists, canmatch, exec, matchmore) all use the same
prototype, which makes the code a bit cleaner.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@16558 65c4cc65-6c06-0410-ace0-fbb531ad65f3
ast_walk_indications(), to walk through the list of indications.
The new method returns an unlocked record, which is no different from the
behaviour of other existing methods in indications.c
(i.e. they all need to be fixed, with refcounts or some similar
method).
Note that ast_walk_indications() uses the pointer argument only as a
search key, so its implementation is completely safe.
In turn, this change allows the removal of AST_MUTEX_DEFINE_EXPORTED.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@16532 65c4cc65-6c06-0410-ace0-fbb531ad65f3
#define DEFAULT_SAMPLE_RATE 8000
#define DEFAULT_SAMPLES_PER_MS ((DEFAULT_SAMPLE_RATE)/1000)
to the main header, and remove equivalent ones from plc.[ch]
This will simplify the cleanup of the codec/ and formats/ files.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@15969 65c4cc65-6c06-0410-ace0-fbb531ad65f3
which is the basis for several simplifications and fixes
to the CLI interfaces.
The core is in cli.c, some documentation on a new function
to help command completion is in cli.h, and one line of
glue code in the other two files.
Next step is to bring in the patches described in #6066 and
other simplifications.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@15817 65c4cc65-6c06-0410-ace0-fbb531ad65f3
and updated to today's version.
The core of the patch is only two files, loader.c
and include/asterisk/module.h, with the other files
touched only to adapt non-standard usages of the
reference counts and localuser lists.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@15784 65c4cc65-6c06-0410-ace0-fbb531ad65f3
writing of common constructs like
chan = !ast_strlen_zero(cdr->channel) ? cdr->channel : "<unknown>";
(or the if/then/else form) into
chan = S_OR(cdr->channel, "<unknown>");
The name can be changed if we find a better (and not too long) one;
currently, it is S as String, OR as it mimics the behaviour of
the || operator, but applied to strings.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@14747 65c4cc65-6c06-0410-ace0-fbb531ad65f3
can override them.
On passing, fix a potential problem in the top level Makefile:
if a static library is not referenced by any of the core objects,
it is not linked in the main program, and will not be available
to modules, which leads to failure at runtime when the modules
are loaded.
This is the case of stdtime/localtime.o, which supplies some core
symbolx, but is only linked in as a library. Fix the problem by
linking in the object.
NOTE: this is intended as a temporary aid to replace the
existing say.c with a newer implementation. Once the
task is completed, we may decide whether or not the ast_say*()
functions should be pluggable or not and possibly revert
part of this change.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@14382 65c4cc65-6c06-0410-ace0-fbb531ad65f3
In this specific case the problem triggered on app_amd.c,
but it keeps coming out from time to time so it is better
to fix it in a more central place.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@14320 65c4cc65-6c06-0410-ace0-fbb531ad65f3
it doesn't report that all allocations are coming from utils.h. Also, add some
more information to the error message astmm reports when a memory allocation
failure occurs.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@13513 65c4cc65-6c06-0410-ace0-fbb531ad65f3
representation into pbx.c so that every file that includes pbx.h does not
unnecessarily get a copy of it
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@13357 65c4cc65-6c06-0410-ace0-fbb531ad65f3
was improperly managed when doing removals or insertions.
also solved issues with app_voicemail init. and reload
solves bug #6557
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@10766 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Rename and export ast_complete_channels for use by cli completion functions
that want to complete from the list of active channels
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@10287 65c4cc65-6c06-0410-ace0-fbb531ad65f3