mirrors/pipewire
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/pipewire/pipewire.git synced 2025-11-05 13:30:02 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions
12134 commits 5 branches 148 tags 50 MiB
Commit graph

135 commits

Author SHA1 Message Date
Wim Taymans
955dde5272 alsa-plugin: use hw_avail in _delay for playback 
for playback streams we want to include the hw_avail, which is the amount of
data that the hardware can read or the amount of data the application has
written.

This is in contrast to using _avail for the capture stream, which is
what the application can read. hw_avail for a capture stream is how many
samples the hardware can write or the amount of free space.

See #1697
 2021-11-24 13:12:23 +01:00
Wim Taymans
743cced63e alsa-plugin: improve debug 2021-11-24 12:40:02 +01:00
Wim Taymans
301e729864 alsa-plugin: fix delay reporting 
The delay should be calculated based on the amount of samples available
to the application, not the hardware.

This overreported the delay.

See #1697
 2021-11-24 12:24:29 +01:00
Wim Taymans
d8e24fca1c alsa-plugin: don't rate match for capture streams 
For capture streams we can simple copy the complete pipewire buffer into
the alsa pcm ringbuffer to make it available for the application.
 2021-11-24 12:24:29 +01:00
Wim Taymans
33f087d4db alsa-plugin: don't clamp elapsed time to stream delay 
There is no reason to limit the elapsed time to the stream delay, it is
possible that it is larger and then it will be clamped later.
 2021-11-24 12:24:29 +01:00
Wim Taymans
6f0c5b6428 alsa-plugin: improve debug 2021-11-24 12:24:29 +01:00
Wim Taymans
25ddf0e3cf alsa-plugin: fix a compiler warnings 2021-11-02 17:18:58 +01:00
Martin Geier
1b96142903 alsa: do not ignore rate change when stream is already created 
gstreamer alsasink can change sample while stream is already created.
In that case, gstreamer calls snd_pcm_ioplug_callback_t::stop,
snd_pcm_ioplug_callback_t::hw_params and snd_pcm_ioplug_callback_t::prepare

Add new flag to the snd_pcm_pipewire_t that is set every time
snd_pcm_pipewire_hw_params is called to prevent using stream with old
sample rate

Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
 2021-11-02 15:37:07 +00:00
Martin Geier
a7d238ed59 alsa: increase precision of delay function 
pw->time.delay is delay in number of frames in pw->time.rate domain,
however snd_pcm_pipewire_delay function is suppose to return number of
frames in io->rate domain. Convert pw->time.delay to io->rate domain to
increase precision when the io->rate is not equal to the pw->time.rate

snd_pcm_pipewire_delay should return how many frames are queued in
pipewire, pw_stream_get_time returns numbers of the queued frames before
snd_pcm_pipewire_process is called, however this function inserts (or
removes) some frames from pipewire. Therefore newly inserted (removed)
frames should be added to pw->time.delay to increase precision.

Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
 2021-11-02 15:37:07 +00:00
Wim Taymans
6ebd589659 alsa-plugin: _prepare will clear the draining state 
Leave the draining state when _prepare is done.

See #1467
 2021-11-02 10:17:31 +01:00
Philippe Normand
cff127a138 pipewire-alsa: Fix build warning 
`pos < 0` always evaluates to false, so remove this condition.
 2021-10-25 07:26:15 +00:00
Wim Taymans
f562b16706 alsa-plugin: use log topics 2021-10-03 08:17:43 +02:00
Jonas Holmberg
61a2577941 pipewire-alsa: fix a memory leak 
Free node_name.
 2021-09-27 11:05:34 +02:00
Wim Taymans
423e8b2f03 Revert "pipewire-alsa: Make it MT safe" 
This reverts commit f22dd9d781.

Obsolete now that pw_init and factory loading/unloading has been
made thread safe.
 2021-09-21 12:35:03 +02:00
Wim Taymans
f6f10669a0 alsa: destroy stream first and then context 
This makes things more deterministic.
 2021-09-15 20:57:10 +02:00
Jonas Holmberg
f22dd9d781 pipewire-alsa: Make it MT safe 
Make pipewire-alsa MT safe so that multi threaded programs can use it
without any extra locks.
 2021-09-08 16:36:58 +02:00
Wim Taymans
7846687df8 alsa: suggest rate to graph with node.rate 2021-08-18 09:56:20 +02:00
Wim Taymans
780f2d645d alsa: enable client-rt for alsa clients 
So that the pipewire thread is running with RT priority.
 2021-08-09 15:28:00 +02:00
George Kiagiadakis
5ee9133b60 alsa plugin: allow specifying a media.role on the virtual device name 
So that we can do:
  aplay -D pipewire:ROLE=Music music.wav
  aplay -D pipewire:ROLE=Notification notice.wav
 2021-06-22 14:48:31 +00:00
Peter Hutterer
0054319d88 meson.build: add -D_GNU_SOURCE to the project arguments 
This appends it to every compilation command so we can get rid of the c_args
for (almost all) executables.
 2021-06-09 07:47:51 +00:00
Peter Hutterer
e0471c6757 pipewire: allow NULL pointers in pw_properties_free() 
Just like the real free() we should just ignore a NULL pointer, makes the
caller code easier for those instances where properties are optional.

Patch generated with concinelle with a few manual fixes.
 2021-06-02 10:56:46 +00:00
Wim Taymans
5a186ece01 alsa: handle plugin errors better 
When the plugins can't be loaded, we fail to make a main_loop.
Handle this case instead of segfaulting.

See #1239
 2021-05-28 11:38:02 +02:00
Peter Hutterer
f3f915e897 pipewire-alsa: fix scan-build errors for unused variables 
All cases of value stored but never read
 2021-05-26 07:51:27 +00:00
Peter Hutterer
7697ed0757 treewide: replace strcmp() == 0 with spa_streq() 
This change is only done in source files for now, header files will be done
separately.
 2021-05-18 22:10:27 +10:00
Wim Taymans
0999f3cf93 alsa-plugin: handle context create error 2021-05-10 13:28:36 +02:00
Peter Hutterer
2405f0942b spa/buffer: rename SPA_MEMBER to SPA_PTROFF 
SPA_MEMBER is misleading, all we're doing here is pointer+offset and a
type-casting the result. Rename to SPA_PTROFF which is more expressive (and
has the same number of characters so we don't need to re-indent).
 2021-05-06 09:39:39 +00:00
Wim Taymans
184bdbeb4c alsa-plugin: Implement query_chmaps again 
Return a list of our supported channel positions.

Fixes #825
 2021-03-04 11:56:28 +01:00
Jonas Holmberg
29be757598 pipewire-alsa: fix drain for capture 
Since pw_stream doesn't call the drained callback after flush for
capture it has to be called explicitly from process.
 2021-03-03 19:00:13 +00:00
Wim Taymans
0aadc0450d alsa: implement poll_descriptors 
In there we can evaluate the poll fd and make sure it blocks or not
in the following poll based on the buffer filled levels. Some API
is very sensitive about this, it seems.

Fixes #433
 2021-03-02 15:03:27 +01:00
Wim Taymans
8386ddff78 alsa-plugin: Fix default channel layout 
We need to use the default channel layout that alsa clients expect.
Remove query chmap, nobody uses that, we need to use the alsa default
channel map.
Implement set_chmap in case anybody wants to set a custom map.
 2021-02-25 17:42:32 +01:00
Wim Taymans
555d5db384 alsa: make sure we are started when draining 
We need to be started before waiting for drain.

Fixes #787
 2021-02-25 12:14:42 +01:00
Wim Taymans
b9241b1d73 alsa: set application.name 2021-02-19 10:06:40 +01:00
Wim Taymans
7053936ffd don't attempt to connect to empty remote name 
Fixes jconvolver startup.

See #744
 2021-02-18 10:22:39 +01:00
Wim Taymans
972e740519 Don't advertize formats we don't yet support 
We don't handle the reverse endian formats yet so make sure we
don't advertize them.
 2021-02-05 12:12:44 +01:00
Wim Taymans
32a3913933 alsa-plugin: use MIN_BUFFERS for max as well 
To get a more reasonable max
 2021-02-04 17:15:46 +01:00
thedannysmith
5e9246e1f7 alsa-plugin: change hw-constrainy buffer-size 2021-02-04 17:14:19 +01:00
Wim Taymans
675a12ef33 alsa: clamp elapsed time to delay 
If the elapsed time is larger than the delay, we are dealing with
an older timestamp.

Fixes #597
 2021-01-20 11:54:19 +01:00
Wim Taymans
fd520c4b39 alsa-plugin: set buffer size correctly 
Our output buffer size is always what is needed to produce one
period of samples. We use nframes from the hardware and the rest
we fill with silence.
 2021-01-18 10:03:19 +01:00
Wim Taymans
d2aee61880 alsa: don't write more than wat is requested 2021-01-17 17:20:59 +01:00
Wim Taymans
b4e3d546e0 alsa: fill frames with silence when underrun 
Keep track of how many samples we need to produce. If we don't have
enough samples in the ringbuffer, fill the rest with silence.

See #564
 2021-01-15 18:34:38 +01:00
Wim Taymans
0c2b4df8e5 alsa: implement _delay to get smoother timestamps 
Fixes #536
 2021-01-09 10:50:39 +01:00
Wim Taymans
3f0cadfa00 alsa: use rate_match to read exactly what we need 
Use the rate match from the resampler to get the exact amount
of samples we need to provide in this cycle.
 2021-01-08 16:55:24 +01:00
Wim Taymans
55c5e48578 alsa: set offset to 0 for playback streams 2020-12-18 15:32:27 +01:00
Wim Taymans
b8300c663b improve some errors 
Silence some -ENOENT warnings, they can happen when the object we
are introspecting is destroyed
 2020-11-29 16:22:50 +01:00
Oschowa
92c541ea03 pipewire-alsa: Make sure to always fill the pipewire buffers with 
silence on process if there is not enough data.
This fixes noise with the retroarch alsa backend on pause, which doesn't
actually pause the device.
 2020-11-01 19:28:38 +01:00
Oschowa
86acbabe8f pipewire-alsa: implement pause 2020-11-01 19:04:37 +01:00
Oschowa
d960724dfc pipewire-alsa: fix inconsistent formating (tabs vs. spaces) 2020-11-01 18:38:46 +01:00
Wim Taymans
e571c0bcb1 alsa: Use "PipeWire ALSA" as the node name 
So that it is easier to know it is the PipeWire one.
 2020-10-27 16:11:14 +01:00
Mersad Jelacic
9de98df1b7 pipewire-alsa: Use monotonic clock 
Use monotonic clock instead of gettimeofday for timestamps.
 2020-10-27 14:53:17 +00:00
Wim Taymans
3567efb7f8 alsa: allow setting playback/capture_node as node name 
See !314
 2020-10-24 15:53:48 +02:00