Alsa UCM device string can contain private configuration prefix required to make
correct device open call. Private prefix is dynamically generated by UCM manager
depending on internal state. Since pulseaudio sink/source port names currently
contain device string, these may change between runs breaking volume database
and module arguments referring to sink/source.
Fix this by skipping UCM private prefix available via `_alibpref` key while
creating UCM mapping name. Mapping object will still contain unmodified
device string for device open call.
See also https://github.com/alsa-project/alsa-ucm-conf/issues/104
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/598>
When child `gsettings-helper` terminates prematurely, unconditionally reading
from child pipe fails in a busy loop until child process is reaped.
Fix this by terminating module upon PA_IO_EVENT_HANGUP or PA_IO_EVENT_ERROR.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/600>
These two log messages are most likely intended for the path that was
just tried, but they are mistakenly printing the name of the port's
current path. Fix them.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/594>
Recently we found an issue of output volume on speaker and headphone,
they should have their own volume but in practice they share one
output volume.
This issue happens on the laptops which use the ucm2 sof-hda-dsp,
originally the speaker has output volume A while the headphone has the
output volume B, suppose the speaker is the active port at the moment
and the output volume is A, users plug a headphone to the jack and the
headphone becomes the active port, in this process, ucm_set_port()
calls _disdev/_enadev which triggers the io_mixer_callback(), in the
meanwhile, the module_device_restore will restore the headphone's
volume to B, it will call set_volume_cb() to set the volume to B, but
this value is not written to hw immediately, during the time of
waiting for the B to be written to the hw, the io_mixer_callback()
calls get_volume_cb(), it reads hw volume and gets the volume A, then
it overrides the output volume to A, this results in the headphone
gets the volume A instead of B.
If a machine doesn't use the ucm, this issue will not happen since the
set_port_cb() will not trigger the io_mixer_callback(). If the ports
don't belong to the same sink/source, this issue also doesn't happen.
BugLink: http://bugs.launchpad.net/bugs/1930188
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/577>
Having G_MESSAGES_DEBUG=all set in the environment (a normal thing to do
when debugging Gnome troubles) causes gsettings-helper to emit a bunch
of helpful gnome debug logs (which is good), but before this change they
were printed on stdout rather than stderr (which was bad!). Rather than
going somewhere the user could see, these log messages were being sent
to the pulesaudio server and interpreted as the src/modules/stdin-util.c
protocol. pulseadio waits to see a '!' message from gsettings-helper
before continuing startup. With the log messages mixed in messing up
the stdin-util protocol, pulseaudio never saw the '!' message, and so
never completed startup.
This simple fix relies on a recent glib > 2.68 (Mar 2021), so builds
against old versions of glib will still have this problem! We consider
this good enough until some complains otherwise.
Fixes: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/1222
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/579>
New card database entry version 5 for card profile is sticky flag.
New messaging API handlers set-profile-sticky and get-profile-sticky.
When card profile is sticky, always restore it even if it is unavailable,
and prevent switching from it when ports become unavailable.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/568>
It seems that in sound context environment variable is not available for
match expression.
This commit utilizes walkthrough to refer to attributes in fw node. The
combination of vendor, model, units is enough to match the node since
the attributes of fw unit doesn't have vendor.
Fix: 37358e42c4 ("alsa: Suppress udev detection of sound card for some units on IEEE 1394 bus")
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/566>
The Volume property on org.bluez.MediaTransport1 is required to utilize
Absolute Volume, but it will only become availabe if the peer device
supports the feature. This happens asynchronously somewhere after the
transport itself has been acquired, after which the callbacks are
attached and software volume is reset.
To prevent race conditions availability of the property is also checked
on startup through a "Get" call.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/239>
Write the current volume to the `Volume` DBus property to keep the
volume on the remote in sync. Without this the remote device shows the
wrong volume, and any attempts to change it will cause an unexpected
jump when the local volume has also been adjusted.
Thanks to prior investments to improve volume synchronization, setting
up callbacks and sending initial volume to the peer for HFP/HSP
implementing this feature is as easy as unconditionally assigning a
valid function to `set_source_volume`. `source_setup_volume_callback`
is already responsible for attaching a `SOURCE_VOLUME_CHANGED` hook and
sending initial (restored) volume to the peer (signifying support for
Absolute Volume - if not derived from the presence of FEATURE_CATEGORY_2
on the profile yet).
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/239>
Like the previous commit this handles `Volume` property changes but
applies them to an A2DP sink instead of source stream. As mentioned in
the AVRCP spec v1.6.2 §5.8 the rendering device (A2DP sink) is
responsible for performing volume attenuation meaning PulseAudio should
pass through audio as-is without performing any attenuation in SW.
Setting a valid pointer to `set_sink_volume` and returning `true` from
`should_attenuate_volume` attaches a hardware callback to `pa_sink` such
that no volume attenuation is performed anymore.
In addition to receiving volume change notifications it is also possible
to control remote volume by writing a new value to the DBus property.
This is especially useful when playing back to in-ear audio devices
which usually lack physical buttons to adjust the final volume on the
sink.
While software volume (used before this patch) is generally fine it is
annoying to crank it up all the way to 100% when a previous connection
to a different device left saved volume on the peer at a low volume.
Providing this bidirectional synchronization is most natural to users
who wish to use physical controls on their headphones, are used to this
from their smartphone, or aforementioned volume mismatches where both PA
as source and the peer as sink/rendering device are performing
attenutation.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/239>
The A2DP spec mandates that the audio rendering device - the device
receiving audio, in our case a `pa_source` - is responsible for
performing attenuation:
AVRCP v1.6.2, §5.8:
The SetAbsoluteVolume command is used to set an absolute volume to be used by the rendering device.
BlueZ models this call as a change of the `Volume` property on the
`org.bluez.MediaTransport1` interface. Supporting Absolute Volume is
optional but BlueZ unconditionally reports feature category 2 in its
profile, mandating support. Hence remote devices (ie. a phone) playing
back audio to a machine running PulseAudio assume volume is to be
changed through SetAbsoluteVolume, without performing any local
attenuation.
Future changes will implement this feature the other way around: setting
an initial value for the `Volume` property as well as propagating
`pa_source` volume changes back to the peer.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/239>
In the case, where the latency is larger than the maximum block size,
module-null-sink will request multiples of the maximum block size from
the sink input instead of limiting the requested amount of data to the
the configured latency.
This patch fixes the problem.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/554>
This reverts commit 96369919e5.
The commit was originally for the issue of Headphone can't output
sound, that was because the Headphone and Lineout share the 1st alsa
mixer and DAC, but this commit introduced a new issue of the speaker
is not muted after switching to headphone.
A recent merged kernel commit (f48652bbe3ae@linux) could fix the 1st
issue, so we could revert the fix of the 1st issue from PA, then the
2nd issue is fixed automatically.
BugLink: https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/747
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/553>
The variable card_name in sink_input_preferred_sink_changed_cb and
source_output_preferred_source_changed_cb could be used uninitialized,
which leads to invalid database entries.
This patch fixes the problem.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/543>
The loopback message may be called after the sink input is already destroyed which causes
a crash. Also memory is leaked because the message object is not correctly freed.
This patch fixes the problems by adding a "dead" flag to the message structure and freeing
the message object on exit.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/541>
When fallback mapping is selected all subsequent profile sets containing
selected mapping are ignored. When there are only e.g. fallback input mappings
available, admitted profile set will only contain one profile with selected
first input fallback mapping and no outputs, and rest of profiles will only
contain outputs and no inputs. When there are only fallback input and output
mappings, there will be no profiles admitted at all.
Fix this by making sure that selected first fallback input or output mapping
is actually allowed to exist in all probed profile sets.
Note while this change allows selected fallback mappings to be found in duplex
configuraitons, probing fallbacks still can fail if there is more than one input
fallback and first one (selected) does not work in duplex configurations.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/539>
A few headsets have issues if HFP HF profile connection is attempted before
HSP HS profile connection is closed. Looks like this could happen because
bluez bluetoothd alows to make simultaneous HSP HS and HFP HF peer connections.
One of affected headsets is WH-1000XM2
Until we find out how to prevent simultaneous HSP HS and HFP HF connections,
when native backend has HFP HF profile enabled (this is the default) do disable
HSP HS completely unless user explicitly request it via discovery modarg.
Do this by adding module-bluetooth-discover arg enable_native_hsp_hs,
default to inverse of enable_native_hfp_hf.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/538>
For mSBC to work correctly the following must be set correctly
- codec object
- transport write method
- transport setsockopt method
Use helper method to set all three simultaneously.
Static configuration structure may be cleaner solution.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/507>
