This is needed for future changes which will require being able to
process the load priority out of order.
Change-Id: Ia23421197f09789940510b03ebbbf3bf24d51bea
* Split off load_dlopen to perform actual dlopen, check results and log
warnings when needed.
* Use flags which minimize number of calls to dlopen required. First
attempt always uses RTLD_GLOBAL when global_symbols_only is enabled,
RTLD_LOCAL when it is not.
This patch significantly reduces the number of dlopen's performed. With
299 modules my system ran dlopen 857 times before this patch, 655 times
after this patch.
Change-Id: Ib2c9903cfddcc01aed3e01c1e7fe4a3fb9af0f8b
This protects the module loader itself against crashing if dlopen is
called on a module from outside loader.c.
* Expand scope of lock inside ast_module_register to include reading of
resource_being_loaded.
* NULL check resource_being_loaded.
* Set resource_being_loaded NULL as soon as dlopen returns. This fixes
some error paths where it was not NULL'ed.
* Create module_destroy function to deduplicate code from
ast_module_unregister and modules_shutdown.
* Resolve leak that occured if a module did not successfully register.
* Simplify checking for successful registration.
Change-Id: I40f07a315e55b92df4fc7faf525ed6d4f396e7d2
Remove nearly all use of regex from ACO users. Still remaining:
* app_confbridge has a legitamate use of option name regex.
* ast_sorcery_object_fields_register is implemented with regex, all
callers use simple prefix based regex. I haven't decided the best
way to fix this in both 13/15 and master.
Change-Id: Ib5ed478218d8a661ace4d2eaaea98b59a897974b
ACO uses regex in many situations where it is completely unneeded. In
some cases this doubles the total processing performed by
aco_process_config.
* Create ACO_IGNORE category type for use in place of skip_category
regex source string.
* Create additional aco_category_op values to allow specifying category
filter using either a single plain string or a NULL terminated array
of plain strings.
* Create ACO_PREFIX to allow matching option names to case insensitive
prefixes.
Change-Id: I66a920dcd8e2b0301f73f968016440a985e72821
We should not do flood detection on video RTP streams. Video RTP streams
are very bursty by nature. They send out a burst of packets to update the
video frame then wait for the next video frame update. Really only audio
streams can be checked for flooding. The others are either bursty or
don't have a set rate.
* Added code to selectively disable packet flood detection for video RTP
streams.
ASTERISK-27440
Change-Id: I78031491a6e75c2d4b1e9c2462dc498fe9880a70
rasterisk does not need to handle setting verbose levels locally, it
should just tell the daemon what it wants and print what it is given.
Just max out the verbose level on the local client so all filtering
happens on the daemon.
ASTERISK-20281 #close
Change-Id: Ia305f75f1fc424a9169bfa30ef70d626ace2c8a8
Using the LIKE operator requires a full table scan of 'astdb', whereas a
comparison operation is able to use the primary key index.
This patch adds a new function to the AstDB API for quick prefix matches
and updates res_sorcery_astdb to utilize it. This showed substantial
performance improvement in my test environment.
Related to ASTERISK~26806, but does not completely resolve it.
Change-Id: I7d37f9ba2aea139dabf2ca72d31fbe34bd9b2fa1
Optimize resource_name_match. This change eliminates use of
ast_strdupa, instead verifying that both basename's are the same length,
then using strncasecmp.
Change-Id: I477275c0e954c99d74be5abfc8bb6545b04e5a3d
There are many places in the code base where we ignore the return value
of fcntl() when getting/setting file descriptior flags. This patch
introduces a convenience function that allows setting or clearing file
descriptor flags and will also log an error on failure for later
analysis.
Change-Id: I8b81901e1b1bd537ca632567cdb408931c6eded7
Duplicate checking was done incorrectly when parsing completion options
from a remote console causing all options to be ignored as duplicates.
Once fixed I had to separate processing of the best match to ensure it
was not identified as a duplicate when it is the only match.
ASTERISK-27465
Change-Id: Ibbdb29f88211742071836c9b3f4d2aa1221cd0f9
The sounds index is rebuilt each time a format is registered or
unregistered. This causes the index to be repeatedly rebuilt during
startup and shutdown.
This patch significantly reduces the work done by delaying sound index
initialization until after modules are loaded. This way a reindex only
occurs if a format module is loaded after startup. We also skip
reindexing when format modules are unloaded during shutdown.
Change-Id: I585fd6ee04200612ab1490dc804f76805f89cf0a
This eliminates some wasteful operations in media_index startup.
* Replace statically set string-fields with char[0].
* Eliminate pointless RAII_VAR's.
* alloc_variant: Avoid pointless ao2_find on new info->variant.
* Stop trying find_variant before alloc_variant.
* process_media_file: replace ast_str with ast_asprintf. This avoids
reallocation of file_id_str.
Overall sounds_index.c is about 27% of Asterisk startup time when using
sample configs. This patch reduces it to 20%. This is a half-fix. The
real problem is that the media_index is regenerated repeatedly - 68
times in my test.
Change-Id: Ia50b752f8efb356f852b05c4be495a6631af8652
* Added start DTMF transfer verbose messages.
* Made associated transfer messages use a similar message format.
* Adjusted message verbose level as requested by initial reporter.
ASTERISK-27449
Change-Id: I2045714586414b3c5ef1f3cc56c1c4af4b31f551
* Add the channel name to diagnostic messages so you will know which
channel failed to transfer.
* Promoted some debug messages to verbose 4 messages.
ASTERISK-27449 #close
Change-Id: Idac66b7628c99379cc9269158377fd87dc97a880
ast_category_get() has an (undocumented) implementation detail where it
tries to match the category name first by an explicit pointer comparison
and if that fails falls back to a normal match.
When initially building an ast_config during ast_config_load, this
pointer comparison can never succeed, but we will end up iterating all
categories twice. As the number of categories using a template
increases, this dual looping becomes quite expensive. So we pass a flag
to category_get_sep() indicating if a pointer match is even possible
before trying to do so, saving us a full pass over the list of current
categories.
In my tests, loading a file with 3 template categories and 12000
additional categories that use those 3 templates (this file configures
4000 PJSIP endpoints with AOR & Auth) takes 1.2 seconds. After this
change, that drops to 22ms.
Change-Id: I59b95f288e11eb6bb34f31ce4cc772136b275e4a
When starting Asterisk in the foreground, there is a perceptible delay
when loading modules that use the ACO and sorcery config frameworks.
For example, a lightly configured res_pjsip took 853ms to load on my
VM.
I tracked down the slowness to the XPath queries used to associate the
relevant documentation with the config options. One improvement was
adding a call to xmlXPathOrderDocElems after loading an XML document.
From the libxml2 docs:
Call this routine to speed up XPath computation on static documents.
The second change was to remove recursive descent and wildcard
operators from the XPath queries. After these changes, res_pjsip takes
85ms to load on my VM and there is no longer a perceptible delay when
starting Asterisk in the foreground.
Change-Id: I45d457f1580e26bf5a2b0dab16e8e9ae46dcbd82
When a format has no pre-recorded sound files, Asterisk has to transcode between
formats. For this, Asterisk has a fixed translation table. If the pre-recorded
sound files are not available in the same sample rate, Asterisk has not only to
transcode but also to resample.
Asterisk has pre-recorded files for SLN (8000 kHz) and SLN16 (16000 kHz).
However before this change, Asterisk did not take the sample rate into account,
because the translation paths to SLN and SLN16 got the same score/weight in the
table. Consequently, you might have got narrow-band audio with siren14, speex32,
silk24, and silk12 although those are (ultra) wide-band audio codecs.
With this change, the distance in sample-rates is taken into account. Now on the
Command-Line interface (CLI) 'core show channels', you should see:
(slin@16000)->(slin@32000)->(speex@32000).
ASTERISK-23735
Reported by: Richard Kenner
Change-Id: I9448295c1978be26f8633b6066395e7bbbe2e213
* Stop using "_COMMAND NUMMATCHES" on remote consoles. Using this
command had doubled the amount of work needed from the Asterisk
daemon for each completion request.
* Fix code formatting.
* Remove static buffer used to send the command, use the same buffer
that will receive the results.
* Move sort from ast_cli_display_match_list.
Change-Id: Ie2211b519a3d4bec45bf46e0095bdd01d384cb69
This rewrites ast_el_strtoarr to use vector's internally, but still
return the original NULL terminated array of strings.
Change-Id: Ibfe776cbe14f750effa9ca360930acaccc02e957
* Stop estimating line count, just print until we run out of matches.
* Stop freeing entries, the caller does that anyways.
* Stop calculating / returning numoutput, it was ignored.
Change-Id: I7f92afa8bea92241a95227587367424c8c32a5cb
Some completion generators are very inefficent due to the way CLI
requests matches one at a time. ast_cli_completion_add can be called
multiple times during one invokation of a CLI generator to add all
results without having to reinitialize the search state for each match.
Change-Id: I73d26d270bbbe1e3e6390799cfc1b639e39cceec
The ability to add to localized storage cannot be supported by
ast_cli_generator. The only calls to ast_cli_generator should be by
functions that need to proxy the CLI generator, for example 'cli check
permissions' or 'core show help'.
* ast_cli_generatornummatches now retrieves the vector of matches and
reports the number of elements (not including 'best' match).
* test_substitution retrieves and iterates the vector.
Change-Id: I8cd6b93905363cf7a33a2d2b0e2a8f8446d9f248
