Commit Graph

82 Commits

Author SHA1 Message Date
wm4 41101c2996 win32: revert wchar_t changes
Revert "win32: more wchar_t -> WCHAR replacements"
Revert "win32: replace wchar_t with WCHAR"

Doing a "partial" port of this makes no sense anymore from my
perspective. Revert the changes, as they're confusing without
context, maintenance, and progress. These changes were a bit
premature anyway, and might actually cause other issues
(locale neutrality etc. as it was pointed out).
2015-08-01 21:09:11 +02:00
wm4 0b52ac8a78 win32: replace wchar_t with WCHAR
WCHAR is more portable. While at least MinGW, Cygwin, and MSVC actually
use 16 bit wchar_t, Midipix will have 32 bit wchar_t. In that context,
using WCHAR instead is more portable.

This affects only non-MinGW parts, so not all uses of wchar_t need to
be changed. For example, terminal-win.c won't be used on Midipix at
all. (Most of io.c won't either, so the search & replace here is more
than necessary, but also not harmful.)

(Midipix is not useable yet, so this is just preparation.)
2015-07-29 00:01:32 +02:00
wm4 d4aaf29a05 ao_wasapi: fix crash on hotplug init error
On init error, the mp_msg macros are actually called. They could cause
a crash because state->log was NULL.
2015-06-17 13:42:31 +02:00
Kevin Mitchell 642f84f922 ao/wasapi: use atomic state variable instead of different events
Unfortunately, because we have proxy objects (pAudioVolumeProxy,
pEndpointVolumeProxy, pSessionControlProxy) it looks like we still
have to use MsgWaitForMultipleObjects and watch for and dispatch
pending messages:

https://msdn.microsoft.com/en-us/library/windows/desktop/ms680112%28v=vs.85%29.aspx
2015-04-04 16:31:14 -07:00
Kevin Mitchell bf3e0bc1da ao_wasapi: code formatting and alignment 2015-04-03 15:40:01 -07:00
Kevin Mitchell 07671ac57b ao_wasapi: passthrough rework
* unify passthrough and pcm exclusive mode format setting/testing
* set passthrough format parameters correctly
* support all of mpv's existing passthrough formats
* automatically test passthrough with exclusive mode and enable
  exclusive if it succeeds, even if it was not explictly requested.
  this obviates the need for --ao=wasapi,wasapi=exclusive
* if passthrough fails (such as the device doesn't support the
  format), fallback to either exclusive pcm or shared mode depending
  on what the user specified. Right now this isn't very useful as
  it still fails due to the decoder path remainin stuck on spdif.

fixes #1742
2015-04-03 15:39:51 -07:00
Kevin Mitchell 4987c1906d ao_wasapi: abstract HRESULT_to_str 2015-04-01 02:30:19 -07:00
Kevin Mitchell e408dd20c7 ao_wasapi: remove redundant casts 2015-03-31 14:13:58 -07:00
Kevin Mitchell b6c28dd26b ao_wasapi: simplify hotplug
Take advantage of the fact that list_devs is called with a
hotplug_inited ao. Also eliminate unnecessary nested function
abstraction of hotplug_(un)init and list_devs. However, keep list_devs
in ao_wasapi_utils.c since it uses the private functions get_device_id,
get_device_name and exposing these would require including headers for
IMMDevice in ao_wasapi_utils.h.
2015-03-31 13:43:32 -07:00
Kevin Mitchell ea00fe0eeb ao_wasapi: fix device listing
remove depricated and convoluted validation. refer instead to the
--audio-device option.
2015-03-31 12:28:41 -07:00
Kevin Mitchell a6bf38bcad ao/wasapi: add ao hotplug
Create a second copy of the change_notify structure for the hotplug
ao. change_notify->is_hotplug distinguishes the hotplug version from
the regular one monitoring the currently playing ao. Also make the
change notification less verbose now that there might be two of them around.
2015-03-31 02:02:54 -07:00
Kevin Mitchell 41c10c3ec2 ao/wasapi: rewrite format search
More clearly separate the exclusive and shared mode format discovery.
Make the exclusive mode search more systematic in particular about
channel maps (i.e., use chmap_sel). Assume that the same sample format
/ sample rates work for all channels to narrow the search space.
2015-03-26 05:33:57 -07:00
Kevin Mitchell c52833bf16 ao/wasapi: move resume to audio thread
This echanges the two events hForceFeed/hFeedDone for hResume. This
like the last commit makes things more deterministic.

Importantly, the forcefeed is only done if there is not already a full
buffer yet to be played by the device. This should fix some of the
problems with exclusive mode.

This commit also removes the necessity to have a proxy to the
AudioClient object in the main thread.

fixes #1529
2015-02-23 14:02:08 -08:00
Kevin Mitchell 446fd5a43a ao_wasapi: move reset into audio thread
This makes things a bit more deterministic. It ensures that the audio
thread isn't doing anything between IAudioClient_Stop(),
IAudioClient_Reset() and setting the sample_count to 0.

Buffer overfilling on resume is still a problem in exclusive mode (see
next commit).
2015-02-23 14:01:05 -08:00
wm4 f61b8b312d win32: request UTF-16 API variants, Vista+ APIs, and COM C macros
Put the Vista+ (_WIN32_WINNT) and the COM C (COBJMACROS) defines into
the build system, instead of defining them over and over in the code.
2015-01-07 21:42:44 +01:00
Kevin Mitchell 6a6620a554 ao/wasapi: style/code formatting tweaks 2015-01-02 14:50:59 -08:00
Kevin Mitchell b83e447e2b ao/wasapi: remove unnecessary check of audio thread input
it would have caused a deadlock if it fired anyway.
2014-11-28 10:52:48 -08:00
Kevin Mitchell 8908b80b77 ao/wasapi: more missed cleanup on failure 2014-11-28 10:52:43 -08:00
Kevin Mitchell 7b5baf3b53 ao/wasapi: check return values
Only issue a warning for failure of wasapi_enumerate_devices and
wasapi_fill_VistaBlob.
2014-11-28 10:52:12 -08:00
Kevin Mitchell 3b0b89c0af ao/wasapi: just return 0 unconditionally from the thread
We weren't actually checking this value anyway. We only really
cared about init failure, which was checked another way.
2014-11-28 10:48:36 -08:00
Kevin Mitchell d40c0d4f90 ao/wasapi: fix unmatched CoUninitialize() on init failure 2014-11-28 10:48:36 -08:00
Kevin Mitchell d0c37f0731 ao/wasapi: more debugging messages 2014-11-28 10:48:07 -08:00
Kevin Mitchell e4aaaf69df ao/wasapi: make get_device_delay return an error code 2014-11-28 08:02:16 -08:00
Kevin Mitchell c545c406fa ao/wasapi: increase buffer size to 50 ms
Before it was the default device period, which was too small
causing glitches on on entering/exiting fullscreen.
2014-11-18 05:03:33 -08:00
Kevin Mitchell 22bf0a78df ao/wasapi: Don't free stuff the thread may still be using on timeout
In the unlikely event of a timeout waiting for the audio thread to return,
don't free stuff that it may still be using.
2014-11-17 23:46:38 -08:00
Kevin Mitchell 20d42b3475 ao/wasapi: also free the threadLoop handle on uninit
http://msdn.microsoft.com/en-us/library/windows/desktop/ms682453%28v=vs.85%29.aspx
2014-11-17 23:43:51 -08:00
Kevin Mitchell 23f52fd41b ao/wasapi: fix leaked event handles 2014-11-17 23:32:28 -08:00
Kevin Mitchell ebd161b256 ao/wasapi: fix race condition in uninit on failure.
When the audio thread fails to properly init, it signals failure
to the main thread, AND THEN starts to clean up. For this to work,
ao_init callback must not return until the thread's cleanup is finished.
This is correctly handled in the ao_uninit callback by waiting for
the thread to exit, so just call that to clean up the main thread.
I have no idea why I didn't do this in the first place.
2014-11-17 23:32:13 -08:00
Kevin Mitchell 4c8b841fc4 ao/wasapi: request ao reload on thread_feed failures
Even with change notifications, there are still (rare) cases when the
feed thread gets AUDCLIENT_DEVICE_INVALIDATED. So handle failures in
thread_feed by requesting ao_reload.
2014-11-17 04:31:22 -08:00
Kevin Mitchell 6c512892d4 ao/wasapi: request reset on appropriate events
on changes to PKEY_AudioEngine_DeviceFormat, device status, and default device.
call ao_reload directly in the change_notify "methods".

this requires keeping a device enumerator around for the duration of
execution, rather than just for initially querying devices
2014-11-17 04:31:20 -08:00
Kevin Mitchell 3da6f723c6 ao/wasapi: tidy up better on failure
Before, failures, particularly in the thread loop init, could lead to a
bad state for the duration of mpvs execution. Make sure that
everything that was initialized gets properly and safely
uninitialized.
2014-11-17 04:13:52 -08:00
Kevin Mitchell e28102f1a8 ao/wasapi: improve error messages and add more debug statements
also enforce more consistency in the exit codes and error handling

thanks to Jonathan Yong <10walls@gmail.com>
2014-11-17 04:13:49 -08:00
Kevin Mitchell d4393be0f9 ao/wasapi: make calling of thread_init consistent with thread_uninit 2014-11-17 03:37:07 -08:00
Kevin Mitchell 6eb5c6d186 ao/wasapi: reenable the reset function
the race condition that necessitated disabling
this was fixed in
e440352313
2014-11-17 03:37:07 -08:00
wm4 e9b0a61444 ao_wasapi: implement device listing 2014-10-13 18:21:45 +02:00
wm4 439a05d8c3 audio/out: remove old things
Remove the unnecessary indirection through ao fields.

Also fix the inverted result of AOCONTROL_HAS_TEMP_VOLUME. Hopefully the
change is equivalent. But actually, it looks like the old code did it
wrong.
2014-09-06 02:30:57 +02:00
wm4 c36faf8c49 audio/out/pull: remove race conditions
There were subtle and minor race conditions in the pull.c code, and AOs
using it (jack, portaudio, sdl, wasapi). Attempt to remove these.

There was at least a race condition in the ao_reset() implementation:
mp_ring_reset() was called concurrently to the audio callback. While the
ringbuffer uses atomics to allow concurrent access, the reset function
wasn't concurrency-safe (and can't easily be made to).

Fix this by stopping the audio callback before doing a reset. After
that, we can do anything without needing synchronization. The callback
is resumed when resuming playback at a later point.

Don't call driver->pause, and make driver->resume and driver->reset
start/stop the audio callback. In the initial state, the audio callback
must be disabled.

JackAudio of course is different. Maybe there is no way to suspend the
audio callback without "disconnecting" it (what jack_deactivate() would
do), so I'm not trying my luck, and implemented a really bad hack doing
active waiting until we get the audio callback into a state where it
won't interfere. Once the callback goes from AO_STATE_WAIT to NONE, we
can be sure that the callback doesn't access the ringbuffer or anything
else anymore. Since both sched_yield() and pthread_yield() apparently
are not always available, use mp_sleep_us(1) to avoid burning CPU during
active waiting.

The ao_jack.c change also removes a race condition: apparently we didn't
initialize _all_ ao fields before starting the audio callback.

In ao_wasapi.c, I'm not sure whether reset really waits for the audio
callback to return. Kovensky says it's not guaranteed, so disable the
reset callback - for now the behavior of ao_wasapi.c is like with
ao_jack.c, and active waiting is used to deal with the audio callback.
2014-05-29 02:24:17 +02:00
wm4 8e7cf4bc99 atomics: switch to C11 stdatomic.h
In my opinion, we shouldn't use atomics at all, but ok.

This switches the mpv code to use C11 stdatomic.h, and for compilers
that don't support stdatomic.h yet, we emulate the subset used by mpv
using the builtins commonly provided by gcc and clang.

This supersedes an earlier similar attempt by Kovensky. That attempt
unfortunately relied on a big copypasted freebsd header (which also
depended on much more highly compiler-specific functionality, defined
reserved symbols, etc.), so it had to be NIH'ed.

Some issues:
- C11 says default initialization of atomics "produces a valid state",
  but it's not sure whether the stored value is really 0. But we rely on
  this.
- I'm pretty sure our use of the __atomic... builtins is/was incorrect.
  We don't use atomic load/store intrinsics, and access stuff directly.
- Our wrapper actually does stricter typechecking than the stdatomic.h
  implementation by gcc 4.9. We make the atomic types incompatible with
  normal types by wrapping them into structs. (The FreeBSD wrapper does
  the same.)
- I couldn't test on MinGW.
2014-05-21 02:21:18 +02:00
Diogo Franco (Kovensky) a0347e0651 ao_wasapi: Use the character set conversion functions from io.h
...rather than rolling out our own. The only possible advantage is that
the "custom" ones didn't use talloc.
2014-03-11 16:37:22 -03:00
Diogo Franco (Kovensky) c5012946ee ao_wasapi: Implement AOCONTROL_UPDATE_STREAM_TITLE 2014-03-11 16:37:22 -03:00
Diogo Franco (Kovensky) f8bdada77f ao_wasapi: Implement per-application mixing
The volume controls in mpv now affect the session's volume (the
application's volume in the mixer). Since we do not request a
non-persistent session, the volume and mute status persist across mpv
invocations and system reboots.

In exclusive mode, WASAPI doesn't have access to a mixer so the endpoint
(sound card)'s master volume is modified instead. Since by definition
mpv is the only thing outputting audio in exclusive mode, this causes no
conflict, and ao_wasapi restores the last user-set volume when it's
uninitialized.
2014-03-11 16:37:21 -03:00
Diogo Franco (Kovensky) f3e9b94622 ao_wasapi: Move non-critical code outside of the event thread
Due to the COM Single-Threaded Apartment model, the thread owning the
objects will still do all the actual method calls (in the form of
message dispatches), but at least this will be COM's problem rather than
having to set up several handles and adding extra code to the event
thread.

Since the event thread still needs to own the WASAPI handles to avoid
waiting on another thread to dispatch the messages, the init and uninit
code still has to run in the thread.

This also removes a broken drain implementation and removes unused
headers from each of the files split from the original ao_wasapi.c.
2014-03-11 16:37:02 -03:00
Diogo Franco (Kovensky) 58011810e5 ao_wasapi: Split into 2 files
ao_wasapi.c was almost entirely init code mixed with option code and
occasionally actual audio handling code. Split most things to
ao_wasapi_utils.c and keep the audio handling code in ao_wasapi.c.
2014-03-11 16:37:02 -03:00
Diogo Franco (Kovensky) f3514fb4bd ao_wasapi: Initial conversion to the new pull model
Gets rid of the internal ring buffer and get_buffer. Corrects an
implementation error in thread_reset.

There is still a possible race condition on reset, and a few refactors
left to do. If feasible, the thread that handles everything
WASAPI-related will be made to only handle feed events.
2014-03-11 16:37:01 -03:00
Diogo Franco (Kovensky) 5c9c81efcc ao_wasapi: Use double math for QueryPerformanceCounter correction
The uint64_t math would cause overflow at long enough system uptimes
(...such as 3 days), and any precision error given by the double math will
be under one milisecond.
2014-03-09 17:56:29 -03:00
wm4 e16c91d07a audio/out: make draining a separate operation
Until now, this was always conflated with uninit. This was ugly, and
also many AOs emulated this manually (or just ignored it). Make draining
an explicit operation, so AOs which support it can provide it, and for
all others generic code will emulate it.

For ao_wasapi, we keep it simple and basically disable the internal
draining implementation (maybe it should be restored later).

Tested on Linux only.
2014-03-09 01:27:41 +01:00
wm4 41f2b26d11 audio/out: make ao struct opaque
We want to move the AO to its own thread. There's no technical reason
for making the ao struct opaque to do this. But it helps us sleep at
night, because we can control access to shared state better.
2014-03-09 00:19:31 +01:00
Diogo Franco (Kovensky) fe03981bbc ao_wasapi: Slightly improve timer accuracy
Use QueryPerformanceCounter to improve the accuracy of
IAudioClock::GetPosition.

While this is mainly for "realtime correctness" (usually the delay is a
single sample or less), there are cases where IAudioClock::GetPosition
takes a long time to return from its call (though the documentation doesn't
define what a "long time" is), so correcting its value might be important in
case the documented possible delay happens.
2014-03-06 17:21:34 -03:00
Diogo Franco (Kovensky) 1d096f9f1b ao_wasapi: Add device latency to get_delay
The lack of device latency made get_delay report latencies shorter than
they should; on systems with fast enough drivers, the delay is not
perceptible, but high enough invisible delays would cause desyncs.

I'm not yet completely sure whether this is 100% accurate, there are
some issues involved when repeatedly pausing+unpausing (the delay might
jump around by several dozen miliseconds), but seeking seems to be
working correctly now.
2014-03-06 17:21:33 -03:00
James Ross-Gowan d26ee98fa6 w32: use safe DLL search paths everywhere
Windows applications that use LoadLibrary are vulnerable to DLL
preloading attacks if a malicious DLL with the same name as a system DLL
is placed in the current directory. mpv had some code to avoid this in
ao_wasapi.c. This commit just moves it to main.c, since there's no
reason it can't be used process-wide.

This change can affect how plugins are loaded in AviSynth, but it
shouldn't be a problem since MPC-HC also does this and it's a very
popular AviSynth client.
2014-01-27 10:04:29 +01:00