13
0
Commit Graph

172 Commits

Author SHA1 Message Date
a9f285b82b debugging timing stats by showing how max values are computed 2021-06-29 09:46:46 -06:00
9d88f370ae debug windows RunLoop timing stats 2021-06-29 09:25:38 -06:00
c2f7b12b0d window compilation fix 2021-06-26 19:02:47 -06:00
fb562ddb61 window compilation fix 2021-06-26 19:02:13 -06:00
74538b7114 window compilation fix 2021-06-26 18:58:23 -06:00
a2974b0c0b portaudio: improved compilation error fix 2021-06-26 12:56:49 -06:00
8bc4464309 portaudio: compilation error fix 2021-06-26 12:51:54 -06:00
7e858b9bda portaudio: add standardized DSP statistics to portaudio backend (uncompiled) 2021-06-26 12:32:23 -06:00
John Emmas
a2b1482cad Fix a problem with ASIO buffer sizes on Windows
PortAudio uses what it calls 'default suggested latencies' but in callback streaming mode, they can result in wildly inaccurate buffer sizing (e.g. the user requests a buffer size of 128 but PortAudio actually instructs ASIO to use a much bigger size).

What we do now is to improve PortAudio's suggested latency calculation by basing it on the actual buffer size requested by the user.
2021-03-18 16:12:28 +00:00
634d325e5d
Prevent deadlock when disconnecting
The backend holds `_port_callback_mutex` while disconnecting ports.
In some cases disconnecting a port can drop the last reference
resulting in a port-deletion from the connection handler.
This in turn will eventually aquire the `_port_callback_mutex`
and deadlock.

This is now circumvented by using atomic operations instead of
taking a lock to set the `_port_change_flag`.

The flag is also used to trigger a latency update in some cases,
atomic is preferable to taking a lock to set this flag.

--

Full bt: https://paste.debian.net/1184056/
Short:

#1  in pthread_mutex_lock ()
#2  in ARDOUR::PortEngineSharedImpl::port_connect_add_remove_callback()
#3  in ARDOUR::BackendPort::~BackendPort()
#4  in ARDOUR::DummyPort::~DummyPort()
#6  in ARDOUR::DummyAudioPort::~DummyAudioPort()
#7  in boost::checked_delete<ARDOUR::BackendPort>(ARDOUR::BackendPort*)
#12 in boost::shared_ptr<ARDOUR::ProtoPort>::reset()
#13 in ARDOUR::Port::drop()
#14 in ARDOUR::Port::~Port()
#15 in ARDOUR::AudioPort::~AudioPort()
#17 in ARDOUR::AudioEngine::add_pending_port_deletion(ARDOUR::Port*)
#20 in boost::detail::sp_counted_base::release()
#37 in ARDOUR::PortManager::connect_callback() at libs/ardour/port_manager.cc:788
#38 in ARDOUR::DummyAudioBackend::main_process_thread() at libs/backends/dummy/dummy_audiobackend.cc:1018
2021-02-04 21:06:20 +01:00
21e66216be
Separate user-set pretty-names from hardware I/O names
This allow to restore original engine port-names as set
by the backend. ALSA MIDI, CoreAudio, CoreMIDI and PortAudio
drivers can provide human readable physical port names for
some devices.
2021-01-18 03:18:20 +01:00
a80064981e
NO-OP: Consolidate PortConnect code into shared parent class 2020-09-05 23:05:57 +02:00
92f54b3e98
Add API to query if systemic latency can be measured 2020-06-30 03:34:26 +02:00
09aa0a3d1a
Consolidate code using pthread_attr_setstacksize
This also adds some stack constraint to rt and fallback threads
that didn't have those before (ALSA MIDI for example)
2020-06-06 18:35:44 +02:00
3d166c7789
Consolidate stack-size and priority of rt-threads 2020-04-23 01:28:58 +02:00
78a66b7180 improved port cleanup as backend is destroyed 2020-04-07 23:26:51 -06:00
32734ce53c fix portaudio backend to use shared_ptr 2020-04-07 16:00:22 -06:00
b9cb306e8b use shared_ptr to manage backend port lifetimes (Pulse,ALSA,Dummy,JACK)
JACK is not yet finished.

Changes also include minor reformatting and a spelling correction (latecies to latencies)
2020-04-07 13:23:49 -06:00
6a6b3f2703 fix another iterator type 2020-04-02 19:45:31 -06:00
cb4997f4a4 fix iterator type 2020-04-02 19:45:31 -06:00
af69ac1285 use clear_ports() 2020-04-02 19:45:31 -06:00
6cdc245967 fix header name 2020-04-02 19:45:31 -06:00
7fc5a6c971 initial try (not compiled) of PortAudio backend using PortEngineSharedImpl 2020-04-02 19:45:31 -06:00
a8ff35accb
Reset DSP load on backend re-start and after freewheeling 2020-03-30 16:58:55 +02:00
e0d5c1426c
NO-OP: fix some Wimplicit-fallthrough
gcc can recognize various regexps in comments. Since C++17 provides
[[fallthrough]], using /* fallthrough */ consistently seems
appropriate until we switch to C++17.

see also https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
2019-09-18 17:37:54 +02:00
47a62b2850
Update backend GPL boilerplate and (C) from git log 2019-08-03 15:53:16 +02:00
076e9fddd4
Remove ambiguous API implementation
* PortEngine::available() implementation
* AudioEngine::connected() wrapper

Eventually we may re-introduce PortEngine::available along
with a libardour internal port-engine.
2018-11-28 15:32:40 +01:00
070f370e40 Fix copy/edit in e047b01aa2 2018-10-18 15:07:45 +02:00
e047b01aa2 add new API for retrieving port flags from backend 2018-10-17 14:11:37 -04:00
e39a8b90d9 amend ab3889ff: portaudio backend uses a vector for connections 2017-09-23 01:06:10 +02:00
ab3889fff8 Propagate Latency to backend/system ports
Set accumulated capture-latency for physical-outputs
and accumulated playback-latency for physical-inputs
after Ardour is done setting all non-physical port latencies.

This will be needed for latency-compensation of the complete graph.
2017-09-23 00:34:01 +02:00
30b087ab3d globally change all use of "frame" to refer to audio into "sample".
Generated by tools/f2s. Some hand-editing will be required in a few places to fix up comments related to timecode
and video in order to keep the legible
2017-09-18 12:39:17 -04:00
7258c6334d Namespace PBD::RingBuffer
class RingBuffer<> is a very generic name and should not pollute the
global namespace.
2017-09-16 16:37:29 +02:00
467c801ce8 Abstract definition of rt-scheduler policy
pthread-w32 does not support pthread_setschedparam() with
SCHED_FIFO and bails out. While pthread_create() simply ignores the policy
and sets the priority regadless.

This only affects ctrl-surface event-loops & AutomationWatch on Windows.
2017-08-29 20:35:36 +02:00
df659e91f1 Consolidate _realtime_pthread_create() into libpbd. 2017-08-29 17:25:23 +02:00
26dc287c00 Fix non-buffered PA backend. Wait for backend to become active
This hopefully fixes an issue with port-registration (new session)
being skipped because PortAudioBackend::available() still false
until the first callback.
2017-08-07 23:23:39 +02:00
fd6377a5af PortAudio: skip process callbacks until ports are established 2017-08-07 19:27:37 +02:00
fde07d4f4f amend bc46a7e2e9; fix blind coding typos 2017-08-05 16:17:31 +02:00
bc46a7e2e9 PortAudio backend RT-safe MIDI buffer allocation 2017-08-05 15:36:36 +02:00
2b7c585dba Update backend API: read-only MIDI input buffers 2017-08-05 15:36:36 +02:00
c125cd2647 Reduce DSP thread priority (main-i/o > midi i/o > computation) 2017-08-05 01:28:11 +02:00
6b5d374352 Some more assert() debugging 2017-08-03 23:46:05 +02:00
f9db9bf5fd Retain order of concurrent MIDI events
This fixes an issue with FaderPort8 (and maybe other surfaces or synths).
2017-04-08 23:43:50 +02:00
9bf4f339b9 revert 335debfa for the PA backend
The same device may have different names for input + output
2016-11-27 00:42:08 +01:00
335debfa2f Add API to enforce valid device selection. 2016-11-26 15:31:21 +01:00
d10913e2de Fix intermittent hang when stopping PortaudioBackend
This issue is not always reproducible but when it does occur it happens
somewhat consistently on both 32bit and 64bit builds(Tested on Windows 7).

The midiOutReset call does not return (or it takes so long that it might as
well be indefinite) and as it is not strictly necessary just remove it.

Resolves: http://tracker.ardour.org/view.php?id=7095
2016-11-15 09:36:52 +10:00
8b5fad0cf9 Use better debug output in PortaudioBackend
When failing to open audio stream in callback mode
2016-11-10 15:23:43 +10:00
cf52d6e4b4 enough with umpteen "i18n.h" files. Consolidate on pbd/i18n.h 2016-07-14 14:45:23 -04:00
d8ac5eddf7 add missing include 2016-05-06 21:25:21 +02:00
514da9881d convert WinMME Device names to UTF-8 2016-05-06 20:59:24 +02:00