Allow asynchronous changes in transport state in the sinks/sources.
Also allow transport acquire to be actually synchronous, in this case it
must set transport state during acquire call.
Separate driver start/stop from transport start/stop.
Add some guards against doing processing when there has been an error or
the node is not started. Set error status to IO. Continue driving on IO
errors.
When reading the timerfd gives an error, we should return right away
because the timeout did not happen.
If we change the timerfd timeout before reading it, we can get -EAGAIN.
Don't log an error in that case but wait for the new timeout.
Use separate timers for driving graph and for flushing, since they don't
have the same period.
Flushing is done based on the time positions of the next sample to be
written, so it will stay in sync with the graph. Because writing too
much data to SCO sockets generally only causes the device to skip ahead,
we don't need to handle the case where the writing has been lagging.
This fixes simultaneous playback to both ALSA and SCO sinks from the
same graph, with SCO as driver, which previously produces broken sound
(e.g. with pw-play --latency 512, linked to the two sinks) ALSA nodes
require regular driver intervals, which was not true previously.
The PropInfo either has a registered id (and then also a name from the
type-info) or a custom name as a string.
In all cases, the description contains a free form text that clarifies
the property.
Use the description in the stream controls name.
Remove quantum limitation from a2dp-sink, and adjust how flushing is
done.
The "low-latency" A2DP codecs are not able to flush all data at once, so
for them flush based on a timeout, such that "excess samples" for each
quantum is bounded. We also limit excess samples for the other A2DP
codecs, based on some testing on flaky headset/adapter combinations (for
most cases, this does not appear to matter).
Leave decision of packet sizes to the codecs. Currently, we send packets
based on min_latency, but sendinf full packets might help with stutter
on some headset/adapter combinations. The slightly increased latency
hardly matters against the 100ms delays in BT headsets.
Bump codec API version.
When sink is follower, and no data to write is available, it should not
schedule a timeout, but wait for the driver to wake it up again.
Fixes process ending up busylooping in data thread as follower, under
some conditions.
Also, clean up the code to be more clear about timeout logic. Just loop
directly instead of setting timeout 1, if we need to just flush more
immediately.
Make the alignment parameter optional when negotiating buffers.
Default to a 16 bytes alignment and adjust for the max cpu
alignment.
Remove the useless align buffer parameter in plugins, we always
set it to 16 anyway.
If socket write results to EAGAIN/EWOULDBLOCK or rx data starts late,
spa_bt_sco_io_write may return 0, and we should give up and skip ahead
(and not loop in RT thread with very small timeout).
Mark some structures, arrays static/const at various places.
In some cases this prevents unnecessary initialization
when a function is entered.
All in all, the text segments across all shared
libraries are reduced by about 2 KiB. However,
the total size increases by about 2 KiB as well.
When we add a new listener to an object, it will emit the full state
of the object. For this it temporarily sets the change_mask to all
changes. Restore the previous state after this or else we might not
emit the right change_mask for the next listener.
Consider the case where one there are two listeners on an object.
The object emits a change and the first listener wants to enumerate the
changed params. For this is adds a new listener and then triggers the
enumeration. If we set the change_mask to 0 after adding the listener,
the second listener would get a 0 change_mask and fail to update
its state.
Add media.name so that restore-stream can store this stream.
Remove node.driver because the stream is slaved to the connected ALSA device.
Remove pause-on-idle in all cases, as this is now configurable from
the config file.
This may avoid infinite loops if parameters are being set based on events
sent by parameter changes. It's also what alsa-acp devices do, so bluez5
should follow.
Remove wrong adjustment to the write timeout with mSBC.
The time between writes should be the 7.5ms at which we encode, which is
correct even when the frames get fragmented to incommensurate tx packets
(on USB BT adapters).
Fixes sound on RTL8176B. This adapter has large reported write_mtu, so
the wrong delay affected it.
Move SCO polling to a single place, and abstract mtu handling.
Autodetect suitable tx packet size based on rx, instead of relying on
the kernel providing correct values.
Since the data thread accesses the spa_bt_transport, its destroy event
needs to sync with data thread to avoid races.
Also check transport is present in places that need it.
When reusing buffers, do it in the same way as the other sinks.
The previous code did not set a valid next buffer for the i/o,
possibly causing it to hang.
Backends don't necessarily allow for opening the same device multiple
times, and it shouldn't be necessary.
Since source and sink are not necessarily both running at the same time,
refcount the transport acquire/release so that it knows to close the fd
only when no source/sink is running.
Let the transport manage the fd lifecycle, also closing it once it is
not needed.
Don't return the fd from acquire(), since each transport is associated
with a single socket fd.
Instead of using a hardcoded value for the minimum delay between writes,
select it according to the mtu/packet size that is currently used.
This is necessary for correct playback with lower mtu values.
It seems that some BT adapters don't return correct SCO MTU, see
https://marc.info/?t=148586768600002&r=1&w=2
Do not try to autodetect the MTU and use a fix MTU size of 48 bytes.
