This forms the base for being able to expose all ports of all
profiles (even inactive ones) to clients.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Since both cards and sinks can hold references to a port, it makes
sense to reference count them. Although no current implementation
actually has sinks with ports but without a card, it felt wrong
to make it harder to make such an implementation in the future.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
This makes handling of echo-cancel streams (and potentially other
filters that need to work with multiple streams) cleaner. We do this by
adding the ability to apply filter on a group of streams rather than
always treating streams individually.
This is required, for example, when changing the input source for an
echo-cancel source output. When this happens, we want to change the
module-echo-cancel instance to actually cancel between the new source
and old sink.
To do this when one stream in a group moves, we create a new filter
instance between the current master sink/source pair and move the
relevant streams to the virtual sink/source from this new instance.
This moves out code from module-stream-restore and makes an internal API
out of it to get a "stream group" for a given sink input or source output.
This is factored out for reuse in module-filter-*.
The stream group basically provides some means of attaching a logical
identification to the stream (by role, application id, etc.).
When autoloaded, it is expected that module-filter-apply (or whatever is
loading us) will take care of applying the filter on the correct
sink/source master. Instead of adding complexity by tracking what is
currently being filtered, we just disallow filtering anything except the
original master sink/source and let module-filter-apply or whatever is
loading us deal with dynamic sink/source changes.
This makes what devices are being cancelled clearer in the UI (at the
cost of being somewhat less clear when multiple devices of the same name
are plugged, but at least that's a much smaller set than everyone).
In pa_create_stream_callback, a stream is inserted into
s->context->record_streams only if it's a record stream. Otherwise it's
inserted into s->context->playback_streams. However, in stream_unlink the
stream is removed from s->context->playback_streams only if it's a playback
stream and otherwise it's removed from s->context->record_streams.
Thus, if the stream is an upload stream, we first insert it into
s->context->playback_streams in pa_create_stream_callback and then try to
remove it unsuccessfully from s->context->record_streams in stream_unlink. This
means that we are leaking hashmap entries until the context is freed,
constantly consuming more memory with applications that upload and unload a
large number of samples through one context.
Of course, this begs the question whether upload streams even belong in either
of those hashmaps. I don't want to mess around with the code too much at this
point though, so this patch should be a sufficient improvement.
This adds some infrastructure for canceller implementations to also
perform acoustic gain control. Cancellers now have a couple of new API
calls that allow them to get/set capture volume.
This is made slightly complex by the fact that cancellation happens in
thread context while most volume mangling needs to be done in main
context. To deal with this, while getting the volume we save source
volume updates as they are propagated to thread context and use this
cached value for queries. To set the volume, we send an async message to
main context and let that set the source volume.
This updates corked streams' resamplers when switching sample rates on a
sink/source, which means the restriction of allowing sample rate updates
only when no streams are attached to a sink/source is now relaxed to
preventing updates only when there is a running stream attached.
The purpose of this command is to print all the internal volume
variables for sinks/sources and all corresponding
sink-inputs/source-outputs to make debugging and reasoning about
volume-related issues easier.
This adds a boolean module parameter to disable automatic dynamic
latency readjustments on underruns, but leaves automatic dynamic
watermark readjustments untouched.
Allow module-bluetooth-device to listens to HandsfreeGateway state
changes using DBUS signals. When an handsfree connects, module-bluetooth-device
is loaded and goes to playing state. When the handsfree disconnect audio,
the card profile is set to "off". If the headset connects audio again after
that, the card profile should switch to "hfgw" again to match state of audio
connection.
If card profile is set to "off", the audio stream should be released.
Current implementation releases the stream when the card profile
is changed to "hsp" or "hfgw" again and immediatly reconnects after that.
This happens in the following scenario :
An HandsfreeGateway connects RFCOMM and then SCO. A card appears in
PA and can be used. If for some reason, SCO is disconnected,
module-bluetooth-device is unloaded. The card will disappear, even
if RFCOMM is still connected. After that, it is not possible to
connect SCO again from PA.
This patch will add the necessary quirks so that pulseaudio can register
an endpoint on the /MediaEndpoint/HFPHS path. This endpoint is to be
used for HFP Handsfree profile.
Some tests (remix-test, sig2str) only display information, so they are not
useful for automated testing. Others (interpol-test, once-test, thread-test)
do return an error on failure, so should be included in TESTS.
Instead of spilling thousands of lines of output, make check now runs the
test-suite in about 100 lines or so. If running under make check, the output of
tests is reduced. The MAKE_CHECK environment variable is used for this, so that
when running the test manually, the full output is still shown. Furthermore,
pa_log is used consistently instead of printf, so that all test output goes to
stderr by default. Colored output from make check goes to stdout.
When a test program exits with a nonzero return value (or an assert is hit),
the test is regarded as a FAIL.
This makes `make check` a little more useful.
This dumps out an additional file with each line having a command of the
form:
p <number of playback samples processed>
c <number of capture samples processed>
d <drift as passed to set_drift()>
The test program can be provided this file to "replay" the data exactly
as when it was run live.
The non-drift-compensation path is retained as-is since it is much
simpler.
This adds the ability for echo cancellers to provide their own drift
compensation, and hooks in the appropriate bits to implement this in the
WebRTC canceller.
We do this by introducing an alternative model for the canceller. So
far, the core engine just provided a run() method which was given
blocksize-sized chunks of playback and record samples. The new model has
the engine provide play() and record() methods that can (in theory) be
called by the playback and capture threads. The latter would actually do
the processing required.
In addition to this a set_drift() method may be provided by the
implementation. PA will provide periodic samples of the drift to the
engine. These values need to be aggregated and processed over some time,
since the point values vary quite a bit (but generally fit a linear
regression reasonably accurately). At some point of time, we might move
the actual drift calculation into PA and change the semantics of this
function.
NOTE: This needs further testing before being deemed ready for wider use.
