pa_sink_get_state() and pa_source_get_state() just return the state
variable. We can as well access the state variable directly.
There are no behaviour changes, except that module-virtual-source
accessed the main thread's sink state variable from its push() callback.
I fixed the module so that it uses the thread_info.state variable
instead. Also, the compiler started to complain about comparing a sink
state variable to a source state enum value in protocol-esound.c. The
underlying bug was that a source pointer was assigned to a variable
whose type was a sink pointer (somehow using the pa_source_get_state()
macro confused the compiler enough so that it didn't complain before).
I fixed the variable type.
pa_sink_input_get_state() and pa_source_output_get_state() just return
the state variable. We can as well access the state variable directly.
There are no behaviour changes, except that some filter sources accessed
the main thread's state variable from their push() callbacks. I fixed
them so that they use the thread_info.state variable instead.
The only thing that the drained state was being used for was "pacmd
list-sink-inputs". In all other cases the drained and running states
were treated as equivalent. IMHO, this usage doesn't justify the
complexity that the additional state brings.
This patch was inspired by a bug report[1] that pointed out an error in
an if condition in pa_sink_input_set_state_within_thread(). The buggy
code is now removed altogether.
[1] https://bugs.freedesktop.org/show_bug.cgi?id=106982
When the user manually switches the profile of a bluetooth headset from
"off" to "a2dp_sink", the port availability changes from "unknown" to
"yes", which triggered a recursive profile change in
module-switch-on-port-available. Such recursivity isn't (and possibly
can't) be handled well (that is, PulseAudio crashed), so let's avoid
doing bluetooth profile changes from module-switch-on-port-available
(they're useless anyway).
Bug link: https://bugs.freedesktop.org/show_bug.cgi?id=107044
Now that both backend-native and backend-ofono can coexist and
backend-ofono is always loaded, even on systems without oFono, failing
to register with org.ofono is not necessarily an error.
This lowers the failure message log level from error to info.
A bit hacky approach, but it allows to preserve LFE output position
even in reduced output modes 2.1 and 4.1.
Signed-off-by: Nazar Mokrynskyi <nazar@mokrynskyi.com>
If a sound card doesn't have the "front" device defined for it, we have
to use the "hw" device for stereo. Not so long ago, the analog-stereo
mapping had "hw:%f" in its device-strings and everything worked great,
except that it caused trouble with the Intel HDMI LPE driver that uses
the first "hw" device for HDMI, and we were incorrectly detecting it as
an analog device. That problem was fixed in commit ea3ebd09, which
removed "hw:%f" from analog-stereo and added a new stereo fallback
mapping for "hw".
Now the problem is that if a sound card doesn't have the "front" device
defined for it, and it supports both mono and stereo, only the mono
mapping is used, because the stereo mapping is only a fallback. This
patch makes the mono mapping a fallback too, so the mono mapping is used
only if there's absolutely nothing else that works.
This can cause trouble at least in theory. Maybe someone actually wants
to use mono output on a card that supports both mono and stereo. But
that seems quite unlikely.
Previously, the "avoid-resampling" option of daemon.conf is to make the
daemon try to use the stream sample rate if possible for all sinks or
sources.
This patch applies this option to module-udev-detect and module-alsa-card
as a module argument in order to override the default value of daemon.conf.
As a result, user can use this argument for more fine-grained control.
e.g.) set it false in daemon.conf and set it true for module-udev-detect
or a particular module-alsa-card in default.pa.(or vice versa)
To set it, use "avoid_resampling=true or false" as the module argument.
Signed-off-by: Sangchul Lee <sc11.lee@samsung.com>
There are only stereo and 5.1 output modes supported natively on this
sound card, but with this config more modes like 2.1, 4.0, 4.1 and 5.0
are now exposed. Also profiles list is cleaner now with all profiles
explicitly specified.
Last thing is removed support for microphone on Linux kernels older than
4.3-rc1, which shouldn't be an issue with future version of PulseAudio
likely be installed on newer kernels anyway.
Signed-off-by: Nazar Mokrynskyi <nazar@mokrynskyi.com>
This should make it easier for clients to elevate their audio threads to
real time priority without having to dig through much through specific
system internals.
BlueZ 4 is no longer supported by BlueZ community for a long long time,
also by moving to BlueZ 5 it should make it even more clearer that
BlueZ 4 is no longer an option.
Attempt to use Acquire method if available since it directly returns
the fd in the reply or an error if that the connection could not be
created while Connect offer neither of these and depend on
NewConnection to deliver the fd.
When a new card shows up (during pulseaudio startup or hotplugged),
pulseaudio needs to pick the initial profile for the card. Unavailable
profiles shouldn't be picked, but module-alsa-card sometimes marked
unavailable profiles as available, causing bad initial profile choices.
This patch changes module-alsa-card so that it marks all profiles
unavailable whose all output ports or all input ports are unavailable.
Previously only those profiles were marked as unavailable whose all
ports were unavailable. For example, if a profile contains one sink and
one source, and the sink is unavailable and the source is available,
previously such profile was marked as available, but now it's marked as
unavailable.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=102902
Currently the loopback module uses sample spec and channel map of the
sink by default. It leads to double resample if source and sink sample
specs are different and no rate/format specified in arguments. This
patch causes the source sample spec and channel map to be used by
default.
This is a working implementation of a build with meson. The server,
utils, and most modules build with this, and it is possible to run from
a build tree and play/capture audio on ALSA devices.
There are a number of FIXMEs, of course, and a number of features that
need to be enabled (modules, dependencies, installation, etc.), but this
should provide everything we need to get there relatively quickly.
To use this, install meson (distro package, or mesonbuild.com) and run:
$ cd <pulseaudio src dir>
$ meson <builddir>
$ ninja -C <builddir>
The iec958 output uses device 2 and the iec958 input uses device 0. The
USB configuration in alsa doesn't set up the device numbers correctly,
which is why we need custom configuration in PulseAudio. Ideally this
would be fixed in alsa, but trying to get help for that wasn't
successful.
jack->melem can be null if the jack disappears between probing the card
and the init_jacks() call. I don't know if jacks actually ever disappear
like that (seems unlikely), but this check is in any case needed as long
as init_jacks() has proper handling for the jack disappearing case
(rather than just an assert).
There was a crash report[1] that indicated that card_suspend_changed()
called report_jack_state() with a null melem. I don't know if that was
because the jack actually disappeared, or is there some other bug too.
[1] https://bugs.freedesktop.org/show_bug.cgi?id=104385
The commit "alsa-util: Set ALSA report_delay flag in pa_alsa_safe_delay()"
broke the build on ALSA versions below 1.1.0 because the time stamp
configuration function was introduced in 1.1.0.
This patch makes the usage of snd_pcm_status_set_audio_htstamp_config()
dependent on ALSA version.
The current code does not call snd_pcm_status_set_audio_htstamp_config()
to configure the way timestamps are updated in ALSA. In kernel 4.14 and
above a bug in ALSA has been fixed which changes timmestamp behavior.
This leads to inconsistencies in the delay reporting because the time
stamp no longer reflects the time when the delay was updated if the
ALSA report_delay flag is not set. Therefore latencies are not calculated
correctly.
This patch uses snd_pcm_status_set_audio_htstamp_config() to set the
ALSA report_delay flag to 1 before the call to snd_pcm_status(). With
this, time stamps are updated as expected.
sco_process_render does not unref the memblock when it encounters an error.
This patch fixes the issue. It also changes the return value to 1 in the case
of EAGAIN. Because the data was already rendered and cannot be re-sent, we
have to discard the block.
Because the modified EAGAIN handling prevents the log message about EAGAIN
after POLLOUT from being printed, the log message was moved to
a2dp/sco_process_render().
The rewrite of the thread function does not change functionality much,
most of it is only cleanup, minor bug fixing and documentation work.
This patch also changes the send buffer size for a2dp sink to avoid lags
after temporary connection drops, following the proof-of-concept patch
posted by Dmitry Kalyanov.
Bug-Link: https://bugs.freedesktop.org/show_bug.cgi?id=58746
Additionally the patch changes the fixed latency for HSP playback from 125
to 25 ms. Tests showed that this produces better audio sync, which is
expected as HSP should have smaller latency than A2DP.
module-allow-passthrough has a (necessary) hack to replicate the default
sink selection and format negotiation from sink-input.c. One thing that
got missed in this replication is the possibility that the sink input is
not compatible with the default sink. When this happen, we now exit
gracefully.
Reconfiguration callback should also be set in case of avoiding resampling
option. This patch set the callback for every case because the callback
has already conditions to leave if it is not needed.
Also unnecessary codes of setting alternate sample rate to 0 are removed.
Signed-off-by: Sangchul Lee <sc11.lee@samsung.com>
This was originally planned to be done by paprefs when it starts, but
since the schema is now fully controlled by pulseaudio, it makes sense
to run the conversion from pulseaudio instead.
g_settings_get_child() returns a new GSettings object that needs to be
freed when it's not used any more. This patch collects all the childern
to a GPtrArray and frees them at the end of main(). They can't be freed
earlier, because that would prevent the "changed" signals from being
delivered.
The removed g_signal_connect() call didn't make sense. The callback
expects to be called when individual module groups are changed, not when
the top level object is changed. Also, module_group_callback() expects
user_data to be non-NULL, but here it was set to NULL.
According to the documentation of g_settings_list_children(), the listed
children may be removed at any time, so g_settings_get_child() may
return NULL. This is probably very unlikely to happen in practice, but
it's good to check anyway.
It is confusing if there's a thing named "module" which defines up to 10
modules to load. Calling the thing a "module group" instead should make
it easier to understand.
Originally the idea was to provide the "modules" schema with paprefs,
but since module-gsettings refers to the "modules" schema in its code,
that would make module-gsettings depend on paprefs, which is not good.
Now all schemas are provided by module-gsettings, so the paprefs
dependency is avoided. Unfortunately this means that if paprefs is
modified to load some new modules, the schema in pulseaudio needs to be
updated as well.
GConf is deprecated, and distributions are removing it. paprefs depends
on GConf, so in order to avoid paprefs getting removed as well, paprefs
has to be changed to use something else than GConf. GSettings is the
easiest alternative to migrate to, although it has the same problems
that GConf had: no support for system mode or networking.
This patch takes the non-GConf specific code from module-gconf and puts
it in stdin-util.[ch], which is then reused by module-gsettings.
module-gsettings is designed to be very similar to module-gconf.
Migration is expected to happen as follows: Distributions update
PulseAudio and paprefs at the same time, or first PulseAudio and then
paprefs. paprefs depends on module-gsettings, and module-gsettings
conflicts with module-gconf. Therefore module-gconf gets automatically
removed during the paprefs update. After the update an old PulseAudio is
likely to be running with module-gconf loaded. If the user tries to use
paprefs during this period, whatever the user does in paprefs won't have
any effect until PulseAudio is restarted (probably by a reboot or
relogin). This is not ideal, but will have to do.
When module-gsettings is loaded, it runs gsettings-data-convert
(implemented in a later patch). That will copy the settings from GConf
to GSettings. If gsettings-data-convert is not available (it's part of
GConf, so it may have already been uninstalled), then any previous
paprefs settings are lost.
The alsa sink calls pa_sink_suspend() from the set_port() callback.
pa_sink_suspend() can only be called from the main thread, but the
set_port() callback was often called from the IO thread. That caused an
assertion to be hit in pa_sink_suspend() when switching ports.
Another issue was that pa_sink_suspend() called the set_port() callback,
and if the callback calls pa_sink_suspend() again recursively, nothing
good can be expected from that, so the thread mismatch was not the only
problem.
This patch moves the mixer syncing logic out of pa_sink/source_suspend()
to be handled internally by the alsa sink/source. This removes the
recursive pa_sink_suspend() call. This also removes the need to have the
mixer_dirty flag in pa_sink/source, so the flag and the
pa_sink/source_set_mixer_dirty() functions can be removed.
This patch also changes the threading rules of set_port(). Previously it
was called sometimes from the main thread and sometimes from the IO
thread. Now it's always called from the main thread. When deferred
volumes are used, the alsa sink and source still have to update the
mixer from the IO thread when switching ports, but the thread
synchronization is now handled internally by the alsa sink and source.
The SET_PORT messages are not needed any more and can be removed.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=104761
The suspend cause isn't yet used by any of the callbacks. The alsa sink
and source will use it to sync the mixer when the SESSION suspend cause
is removed. Currently the syncing is done in pa_sink/source_suspend(),
and I want to change that, because pa_sink/source_suspend() shouldn't
have any alsa specific code.
There are no behaviour changes, the code from almost all the SET_STATE
handlers is moved with minimal changes to the newly introduced
set_state_in_io_thread() callback. The only exception is module-tunnel,
which has to call pa_sink_render() after pa_sink.thread_info.state has
been updated. The set_state_in_io_thread() callback is called before
updating that variable, so moving the SET_STATE handler code to the
callback isn't possible.
The purpose of this change is to make it easier to get state change
handling right in modules. Hooking to the SET_STATE messages in modules
required care in calling pa_sink/source_process_msg() at the right time
(or not calling it at all, as was the case on resume failures), and
there were a few bugs (fixed before this patch). Now the core takes care
of ordering things correctly.
Another motivation for this change is that there was some talk about
adding a suspend_cause variable to pa_sink/source.thread_info. The
variable would be updated in the core SET_STATE handler, but that would
not work with the old design, because in case of resume failures modules
didn't call the core message handler.
The previous code made the SET_STATE message fail if trigger() failed.
However, trigger() was called after pa_sink/source_process_msg(), which
meant that the main thread that sent the SET_STATE thought that resuming
failed, but nothing was undone in the IO thread, so in the IO thread
things seemed as if the sink/source was successfully resumed. (I don't
use OSS myself, so I don't know what kind of practical problems this
could cause).
Unless some complex undo logic is implemented, I believe it's best to
ignore all failures in trigger(). Most error cases were already ignored,
and the only one that wasn't ignored doesn't seem too serious.
I also moved trigger() to happen before pa_sink/source_process_msg(),
which made it necessary to add new state parameters to trigger(). The
reason for this move is that I want to move the SET_STATE handler code
into a separate callback, and if things are done both before and after
pa_sink/source_process_msg(), that makes things more complicated.
The previous code checked the return value of
pa_sink/source_process_msg() before calling trigger(), but that was
unnecessary, since pa_sink/source_process_msg() never fails when
processing the SET_STATE messages.
When resuming a sink or source, pa_sink/source_process_msg() should be
called only if resuming is successful. pa_sink/source_process_msg()
updates thread_info.state and notifies streams about the new state, but
if resuming fails, there's no state change.
pa_sink_get_state() is supposed to be used from the main thread. In this
case it doesn't really matter, because the SET_STATE handler is executed
while the main thread is waiting, but since the state is available also
in thread_info, let's use that. All other modules use thread_info.state
too, so at least this change improves consistency.
Also, we can use the PA_SINK_IS_OPENED macro to simplify the code a bit.
