u->asyncmsg is accessed from two IO threads. teardown() shouldn't
flush the queue from the main thread while both IO threads are still
potentially using the queue. This patch fixes that error by flushing
the queue from the sink input thread when the sink input is being
unlinked.
Flushing the queue in teardown() caused this assertion in
pa_asyncmsgq_get() to crash sometimes: pa_assert(!a->current)
process() may be called with a stream that doesn't have its sink/source set.
This can happen if the proplist change callback is called when the stream is
moving.
The sink input may_move_to() callbacks can be called while the source
output is not connected to any source (i.e. is currently moving too),
and vice versa.
Thanks to Frédéric Dalleau for reporting this bug.
ipacl-test fails if there is no SSH server running on your machine.
Since it is not a PulseAudio error not to have an SSH server running,
this test should not be run as part of the "make check" test suite.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Previously, a drain request was acknowledged up to two hw buffers
too late, causing unnecessary delays.
This implements a new chain of events called process_underrun
which triggers exactly when the sink input has finished playing,
so the drain can be acknowledged quicker.
It could later be improved to give better underrun reporting to
clients too.
Tested-by: Dmitri Paduchikh <dpaduchikh@gmail.com>
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
This patch adds support for completion of remote PulseAudio server
arguments it also suppresses error messages when unable to connect to
PulseAudio (only for the completion function).
The previous volume handling could cause ear damage: by default the
ladspa sink volume was 100%, and with flat volumes that would cause
the master sink volume to jump to 100% too.
The previous AAC pass-through patch (commit: 53807e4a) introduced
a new encoding format type: PA_ENCODING_MPEG2_AAC_IEC61937,
which is mostly used in pa_format_info, but forgot to increment the
protocol version number. The version needs to be incremented, because
clients need some way of checking whether the server supports the new
encoding.
If we find a microphone output port, it is probably something else
than a microphone. Therefore label it "Bluetooth output" instead of
"Microphone".
Same goes for Headphones and Speakers, but in the other direction.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
There was a recent thread on Linux Audio Users mailinglist about
whether to do so or not, and it looks like most people would prefer
having a stereo default (but even better would have been a
module-jack-card where you can easily set channels/profiles on the fly).
Reference:
http://lists.linuxaudio.org/pipermail/linux-audio-user/2013-February/091068.html
Reported-by: Kaj Ailomaa <zequence@mousike.me>
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
If there is a proper monitor name, we expose this as a device.product.name
property on the port. This can be useful for UIs who might want to show
this name.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
The alsa mixer kcontrol has "device index" 3, 7, 8, and 9.
We need to configure this properly.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Currently, this function only reads the monitor name, but could
be extended to read e g supported formats as well.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
If BlueZ crashes exactly while PulseAudio waits for the GetProperties reply, the
device has already been removed from the hashmap and therefore an assertion
failure is experienced.
The solution consists of ignoring the reply in these cases.
The problem can be observed in the following traces:
D: [pulseaudio] bluetooth-util.c: Bluetooth daemon appeared.
D: [pulseaudio] bluetooth-util.c: dbus: interface=org.bluez.Manager, path=/, member=AdapterAdded
D: [pulseaudio] bluetooth-util.c: Adapter /org/bluez/497/hci1 created
D: [pulseaudio] bluetooth-util.c: Registering /MediaEndpoint/HFPAG on adapter /org/bluez/497/hci1.
D: [pulseaudio] bluetooth-util.c: Registering /MediaEndpoint/HFPHS on adapter /org/bluez/497/hci1.
D: [pulseaudio] bluetooth-util.c: Registering /MediaEndpoint/A2DPSource on adapter /org/bluez/497/hci1.
D: [pulseaudio] bluetooth-util.c: Registering /MediaEndpoint/A2DPSink on adapter /org/bluez/497/hci1.
D: [pulseaudio] bluetooth-util.c: dbus: interface=org.bluez.Adapter, path=/org/bluez/497/hci1, member=DeviceCreated
D: [pulseaudio] bluetooth-util.c: Device /org/bluez/497/hci1/dev_90_84_0D_B2_C7_04 created
D: [pulseaudio] bluetooth-util.c: dbus: interface=org.freedesktop.DBus, path=/org/freedesktop/DBus, member=NameOwnerChanged
D: [pulseaudio] bluetooth-util.c: Bluetooth daemon disappeared.
E: [pulseaudio] bluetooth-util.c: Assertion 'p->call_data == d' failed at modules/bluetooth/bluetooth-util.c:685, function get_properties_reply(). Aborting.
Enable advanced AEC methods to use different specs (i.e., number of
channels) for rec and out stream. A typical application is beam forming
resp. multi-channel AEC, which takes multiple record channels to produce
an echo-canceled output stream.
This commit alters the EC API as follows: the EC's init() used to get
source and sink's sample spec/channel map. The new interface renamed
source to rec and sink to play and additionally passes sample spec and
channel map of the out stream. The new parameter names of init()
{rec,play,out}_{ss,map} are more intuitive and also resemble to the
parameter names known from run(). Both rec_{ss,map} and out_{ss,map} are
initialized as we knew it from source_{ss,map} before being passed to
init(). The previous EC implementations only require trivial changes,
i.e., setting rec_{ss,map} to out_{ss,map} at the end of init() in case
that out_{ss,map} is modified in init().
The card profile availability flag already provides all the necessary
information and therefore all Bluetooth ports can be merged, leaving
the two generic ones only: "bluetooth-input" and "bluetooth-output". The
availability of these port now represents whether the device is
streaming audio, with the following mapping:
- PA_AVAILABLE_UNKNOWN: some profile connected but not streaming
- PA_AVAILABLE_NO: no profiles connected
- PA_AVAILABLE_YES: some profile streaming (regardless of which)
Each port's flag represents the profiles with the corresponding I/O
capabilities (pa_direction_t).
Use the card profile availability flag instead of port availability in
order to automatically switch profiles, for example when a paired phone
starts streaming A2DP audio.
Use the transport's state to not only update the ports availability, but
also to update the card profile availability flag. The interpretation is
as follows:
- PA_AVAILABLE_UNKNOWN: BT profile is connected but no audio streaming
- PA_AVAILABLE_NO: BT profile disconnected
- PA_AVAILABLE_YES: BT profile connected and audio streaming
Some cards are capable to announce if a specific profile is available or
not, effectively predicting whether a profile switch would fail or would
likely succeed. This can for example be useful for a UI that would gray
out any unavailable profile.
In addition, this information can be useful for internal modules
implementing automatic profile-switching policies, such as
module-switch-on-port-available or module-bluetooth-policy.
In particular, this information is essential when a port is associated
to multiple card profiles and therefore the port availability flag does
not provide enough information. The port "bluetooth-output" falls into
this category, for example, since it doesn't distinguish HSP/HFP from
A2DP.
Generalize the availability flag in order to be used beyond the scope of
ports.
However, pa_port_availability_t is left unchanged to avoid modifying the
protocol and the client API. This should be replaced by pa_available_t
after a validation phase of this new generic enum type.
When the play stream from the EC sink has not enough data available then
the EC implementation is currently bypassed by directly forwarding the
record bytes to the EC source. Since EC implementations maintain their
own buffers and cause certain latencies, a bypass leads to glitches as
the out stream stream jumps forth and back in time. Furthermore, some
EC implementations may also apply noise reduction or other sound
enhancing techniques, which are therefore bypassed, too.
Fix this by passing silence bytes to the EC implementation if the play
stream runs empty. Hence, this patch keeps the EC implementation running
even if the play stream has no data available.
The echo canceller module can pass arguments to the EC implementation
via the module parameter aec_args. However, the echo-cancel-test passes
EC arguments via a separate argv[] option, which is inconsistent. Fix
this.
Unloading modules in the reverse order is the "more logical" thing
to do, and speeds up shutdown somewhat, e g by not loading
module-null-sink at shutdown.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
thread_mq.outq may contain some unprocessed messages, which should be
dispatched before unreffing the sink and source. If the sink and
source are unreffed before all messages to them have been dispatched,
the unreffing won't free the sink and source, and that in turn will
likely cause problems with things getting freed in a wrong order.
A transport should be considered connected only after the connection
procedure is complete, as expressed in audio_state_to_transport_state().
module-bluetooth-device should be loaded only after at least one
transport is not only created (during configuration), but also
connected.
This fixes the issue of premature acquire attempts sometimes experienced
when a headset is connected (issue not present in v3.0 though).
