summaryrefslogtreecommitdiffstats
path: root/Alc
Commit message (Collapse)AuthorAgeFilesLines
* Add a comment about when FreeContext and FreeDevice are calledChris Robinson2011-09-101-3/+3
|
* Trace the device pointer when warning about leftover objectsChris Robinson2011-09-101-3/+3
|
* Set the device-connected flag to false first when handling a disconnectChris Robinson2011-09-101-2/+2
|
* Condense some codeChris Robinson2011-09-101-11/+8
|
* Clean up some code commentsChris Robinson2011-09-101-101/+94
|
* Lock the lists while probing devicesChris Robinson2011-09-101-18/+21
|
* Clean up some commentsChris Robinson2011-09-101-44/+51
|
* Slightly modify a warningChris Robinson2011-09-101-1/+1
|
* Properly destroy contexts that are left on the device when it's closedChris Robinson2011-09-101-37/+46
|
* Avoid locking the device when adding and removing contextsChris Robinson2011-09-101-15/+13
| | | | Also, don't keep track of the number of contexts per device
* Avoid holding the list lock when not neededChris Robinson2011-09-101-118/+172
|
* Warn about contexts that are errantly left currentChris Robinson2011-09-101-0/+11
| | | | | These won't show by default in release mode, unless the log level is changed to show warnings.
* Hold an extra reference on the device for each of its contextsChris Robinson2011-09-101-0/+14
|
* Add some reference counting to the deviceChris Robinson2011-09-101-38/+51
|
* Fix a parameter typeChris Robinson2011-09-101-1/+1
|
* Add an option to trap device errors in a debuggerChris Robinson2011-09-101-0/+25
|
* Add an option to trap context errors in a debuggerChris Robinson2011-09-101-0/+8
|
* Allow PulseAudio's tlength to grow to make sure there's at least two updatesChris Robinson2011-09-091-7/+19
|
* Use an array to get the device format from the user config stringChris Robinson2011-09-091-109/+36
|
* Avoid an unnecessary lockChris Robinson2011-09-021-6/+1
|
* Do an atomic compare-exchange on the global context when destroying a contextChris Robinson2011-09-021-6/+3
|
* Fix multichannel wet gainhf calculationChris Robinson2011-09-021-1/+1
|
* Use an atomic exchange to get the old global context when setting a new oneChris Robinson2011-09-021-2/+3
|
* Log the function name instead of the source file and line numberChris Robinson2011-09-023-19/+15
|
* Trace the reverb buffer sizeChris Robinson2011-09-021-0/+1
|
* Combine the reverb effectsChris Robinson2011-09-011-600/+525
| | | | | | | Updating and processing still differs depending on whether standard or EAX reverb is used or not. The only functional difference should be that the allocated buffer (and subsequent offsets) take into account the modulation and echo times.
* Combine the dedicated (dialog/LFE) effectsChris Robinson2011-09-011-38/+13
|
* Store the filter gains directly in the source instead of duplicate filter ↵Chris Robinson2011-08-312-36/+12
| | | | objects
* Don't set a thread priority if real-time isn't requestedChris Robinson2011-08-311-10/+2
|
* Add a function to initialize a RWLockChris Robinson2011-08-311-5/+10
|
* Rename GetReffedContext to GetContextRefChris Robinson2011-08-301-3/+3
|
* Add a limit to the UIntMap size and use it for sources and effect slotsChris Robinson2011-08-302-10/+17
|
* Use a separate array for the auxiliary slots in the mixerChris Robinson2011-08-302-16/+21
|
* Add functions to retrieve the source from the source map while removing itChris Robinson2011-08-301-0/+29
|
* Use the reference counting functions for the buffer and auxiliary slot refsChris Robinson2011-08-291-1/+1
|
* Use a generic int type to handle enum swapsChris Robinson2011-08-292-5/+5
|
* Make specific functions to lock/unlock UIntMap accessChris Robinson2011-08-292-9/+9
|
* Manually lock the map while iterating through its contentsChris Robinson2011-08-292-0/+9
|
* Use a read-write lock to protect access to the UInt mapsChris Robinson2011-08-291-7/+65
|
* Lock the context as needed for the defer and process callsChris Robinson2011-08-291-2/+26
|
* Hold a reference on the global contextChris Robinson2011-08-291-7/+12
|
* Use atomic exchanges when checking for updates to objects' internal parametersChris Robinson2011-08-291-14/+6
|
* Increment the context reference count when lockingChris Robinson2011-08-281-46/+38
|
* Use a list of contexts in the device instead of an arrayChris Robinson2011-08-282-54/+51
|
* Avoid a global list of contextsChris Robinson2011-08-281-26/+17
|
* Check if the context is valid only if it's replacing the current thread contextChris Robinson2011-08-281-12/+11
|
* Hold a reference on the context when being set as current on the threadChris Robinson2011-08-281-40/+50
| | | | | This should vastly improve efficiency for retrieving the current thread context, as we can be assured the stored context pointer will always be valid.
* Rename ALCcontext_Deref to ALCcontext_DecRefChris Robinson2011-08-281-2/+2
|
* Emulate pthread TLS functions in WindowsChris Robinson2011-08-282-12/+54
|
* Add reference counting to the ALC contextsChris Robinson2011-08-281-25/+33
|