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.
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>
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.
build_pollfd() isn't likely to fail, but if it does, pa_sink/source_put()
will crash on an assertion failure. I haven't seen such crash happening,
this is just something that I noticed while studying the state change
code.
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.
The Intel HDMI LPE driver works in a peculiar way when the HDMI cable is
not plugged in: any written audio is immediately discarded and underrun
is reported. That resulted in an infinite loop, because PulseAudio tried
to keep the buffer filled, which was futile since the written audio was
immediately consumed/discarded.
This patch adds special handling for the LPE driver: if the active port
of the sink is unavailable, the sink suspends itself. A new suspend
cause is added: PA_SUSPEND_UNAVAILABLE.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=100488
Previously max_rewind was always set to the full hw buffer size, but
the actual maximum rewind amount is limited to the part of the hw buffer
that is in use.
The rewind request that was done when lowering the sink latency had to
be moved to happen before updating max_rewind.
The practical benefit of this change: When using a filter source on a
monitor source, the filter source latency is increased by max_rewind.
Without this change the max_rewind of an alsa sink is often
unnecessarily high, which leads to unnecessarily high latency with
filter sources.
Monitor sources themselves don't suffer from the latency issue, because
they use the current sink latency instead of max_rewind for the extra
buffer that they keep to deal with rewinds.
This rejigs the update_rate() logic to encompass changes to the sample
spec as a whole, as well as passthrough status. As a result,
sinks/sources provide a reconfigure() method which allows
reconfiguration as required.
The behaviour itself is currently unchanged -- alsa-sink/-source do not
actually implement anything other than rate updates for now (nor are
they ever requested to). This can be modified in the future, to allow,
for example 24-bit output when incoming media supports it, as well as
channel count changes for passthrough sinks.
Another related change is that passthrough status is now part of
sink/source reconfiguration, and we can stop doing a suspend/unsuspend
when entering/leaving passthrough state. So that part is now divided
in two -- pa_sink_reconfigure() sets the sink in passthrough mode if
required, and pa_sink_enter_passthrough() sets up everything else
(this currently means only volumes, but could disable other processing)
for passthrough mode.
The reported latency of source or sink is based on measured initial conditions.
If the conditions contain an error, the estimated latency values may become negative.
This does not indicate that the latency is indeed negative but can be considered
merely an offset error. The current get_latency_in_thread() calls and the
implementations of the PA_{SINK,SOURCE}_MESSAGE_GET_LATENCY messages truncate negative
latencies because they do not make sense from a physical point of view. In fact,
the values are truncated twice, once in the message handler and a second time in
the pa_{source,sink}_get_latency_within_thread() call itself.
This leads to two problems for the latency controller within module-loopback:
- Truncating leads to discontinuities in the latency reports which then trigger
unwanted end to end latency corrections.
- If a large negative port latency offsets is set, the reported latency is always 0,
making it impossible to control the end to end latency at all.
This patch is a pre-condition for solving these problems.
It adds a new flag to pa_{sink,source}_get_latency_within_thread() to allow
negative return values. Truncating is also removed in all implementations of the
PA_{SINK,SOURCE}_MESSAGE_GET_LATENCY message handlers. The allow_negative flag
is set to false for all calls of pa_{sink,source}_get_latency_within_thread()
except when used within PA_{SINK,SOURCE}_MESSAGE_GET_LATENCY. This means that the
original behavior is not altered in most cases. Only if a positive latency offset
is set and the message returns a negative value, the reported latency is smaller
because the values are not truncated twice.
Additionally let PA_SOURCE_MESSAGE_GET_LATENCY return -pa_sink_get_latency_within_thread()
for monitor sources because the source gets the data before it is played.
If the ALSA device supports granular pointer reporting, we end up in a
situation where we write out a bunch of data, iterate, and then find a
small amount of data available in the buffer (consumed while we were
writing data into the available buffer space). We do this 10 times
before quitting the write loop.
This is inefficient in itself, but can also have wider consequences. For
example, with module-combine-sink, this will end up pushing the same
small chunks to all other devices too.
Given both of these, it just makes sense to not try to write out data
unless a minimum threshold is available. This could potentially be a
fragment, but it's likely most robust to just work with a fraction of
the total available buffer size.
Bug 96741 shows a case where an assertion is hit, because
pa_asyncq_new() failed due to running out of file descriptors.
pa_asyncq_new() is used in only one place (not counting the call in
asyncq-test): pa_asyncmsgq_new(). Now pa_asyncmsgq_new() can fail too,
which requires error handling in many places. One of those places is
pa_thread_mq_init(), which can now fail too, and that needs additional
error handling in many more places. Luckily there weren't any places
where adding better error handling wouldn't have been easy, so there are
many changes in this patch, but they are not complicated.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=96741
If a card has been hot-plugged after pulseaudio start, alsa-lib still has
old configuration in memory, which doesn't have PCM definitions for the
new card. Thus, this error appears, and the device doesn't work:
I: [pulseaudio] (alsa-lib)confmisc.c: Unable to find definition 'cards.USB-Audio.pcm.front.0:CARD=0'
I: [pulseaudio] (alsa-lib)conf.c: function snd_func_refer returned error: No such file or directory
I: [pulseaudio] (alsa-lib)conf.c: Evaluate error: No such file or directory
I: [pulseaudio] (alsa-lib)pcm.c: Unknown PCM front:0
I: [pulseaudio] alsa-util.c: Error opening PCM device front:0: No such file or directory
The snd_config_update_free_global() function makes alsa-lib forget any
cached configuration and reparse all PCM definitions from scratch next
time it is told to open anything.
The trick has been copied from Phonon.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=54029
Signed-off-by: Alexander E. Patrakov <patrakov@gmail.com>
It doesn't work currently (fails and falls back to PCM), due to channel
count mismatch between the sink sample spec and the sample spec required
by IEC61937.
To be reverted when someone implements changing channel count without
switching profiles. This would also be required for HBR passthrough over
HDMI.
Reported-by: Xamindar <junkxamindar@gmail.com>
Signed-off-by: Alexander E. Patrakov <patrakov@gmail.com>
This change doesn't affect behaviour, because accessing boolean fields
in the new data was safe even after the done() call, but it was still
bad style.
FSF addresses used in PA sources are no longer valid and rpmlint
generates numerous warnings during packaging because of this.
This patch changes all FSF addresses to FSF web page according to
the GPL how-to: https://www.gnu.org/licenses/gpl-howto.en.html
Done automatically by sed-ing through sources.
frames_per_block is the mempool's maximum block size in frames
v2 (thanks David Henningson)
* rename max_frames to frames_per_block
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Now that we have switched to using the mixer handle only,
there is no use for sending hctl handles around.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Forcing all mute changes to go through set_mute() makes it easier to
check where the muted field is changed, and it also allows us to have
only one place where notifications for changed mute are sent.
When given an explicit device.description in card_properties, prefer
this information over other default prefixes (e.g. 'Built-in Audio')
when constructing sink/source descriptions.
For example, if I manually configure the card description to be
"FooBar", I then expect that the sinks and created by the card also
have "FooBar" in their description instead of generic "Built-in
Audio".
module-alsa-{sink,source}.c call pa_alsa_{sink,source}_new with
mapping set to NULL. Guard against this, like the rest of the
function does.
module-alsa-card does not use NULL, so this went unnoticed so far.
This allows mappings to override some or all of the sample_spec used to
open the ALSA device. The intention, to start with, is to use this for
devices in UCM that need to be opened at a specific rate (like modem
devices). This can be extended to allow overrides in profile-sets as
well.
All pa_cvolume_snprint(), pa_volume_snprint(),
pa_sw_cvolume_snprint_dB() and pa_sw_volume_snprint_dB() calls have
been replaced with pa_cvolume_snprint_verbose() and
pa_volume_snprint_verbose() calls, making the log output more
informative and the code sometimes simpler.
This patch removes all occurrences of double and triple
newlines.
Command used for this:
find . -type d \( -name ffmpeg \) -prune -o \
-regex '\(.*\.[hc]\|.*\.cc\)' \
-a -not -name 'adrian-aec.*' -a -not \
-name reserve.c -a -not -name 'rtkit.*' \
-exec sed -i -e '/^$/{N;s/^\n$//}' {} \;
Two passes were needed to remove triple newlines.
The excluded files are mirrored files from external sources.
This patch replaces every occurrence of ')\n{' with ') {'.
Command used for this:
find . -type d \( -name ffmpeg \) -prune -o \
-regex '\(.*\.[hc]\|.*\.cc\)' \
-a -not -name core-util.c -a -not \
-name adrian-aec.c -a -not -name g711.c \
-exec sed -i -e '/)$/{N;s/)\n{$/) {/}' {} \;
The excluded files are mirrored files from external sources.
The tsched_watermark is in bytes, not in usecs. Fix this by introducing
a new variable, and also use that variable in some places for optimisation.
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>
Now you can actually see *which* sink/source that sends a specific
message to the log, which is quite useful if you have more than
one sound card.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
I was looking at a log that showed that a suspend happened (at
a strange time), but the log didn't tell me why the suspend was done.
This patch tries to make sure that that won't happen again.
We inadvertantly stopped supporting non-standard rates when the
passthrough work was done. This makes sure that if no standard rates are
supported, we try to fallback to whatever ALSA gives us.
When a rewind is requested on a sink input, the request parameters are
stored in the pa_sink_input struct. The parameters are reset during
rewind processing, and if the sink decides to ignore the rewind
request due to being suspended, stale parameters are left in
pa_sink_input. It's particularly problematic if the rewrite_bytes
parameter is left at -1, because that will prevent all future rewind
processing on that sink input. So, in order to avoid stale parameters,
every rewind request needs to be processed, even if the sink is
suspended.
Reported-by: Uoti Urpala
Refactor code to fetch avail, delay and timestamp values
in a single call to snd_pcm_status().
The information reported is exactly the same as before,
however it is extracted in a more atomic manner to
improve timer-based scheduling.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Sometimes the kernel does not schedule us in due time, thus causing
an underrun. Adding a detection and a debug message will be a helpful
step in determining the cause of an underrun.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
