When the driver node is destroyed (like when unplugging the cable) it
will drop/pause all of the follower ALSA nodes. This is something that
happens internally because of how the ALSA nodes work together, on the
PipeWire level, the nodes are still running and they will just be moved to
another driver.
The problem is that nothing will then start the nodes again after moving
it to the new driver. Fix this by keeping track of the desired target
state of the ALSA node and restoring that state when we detect that we are
paused when moved to a new driver.
Fixes#4401
Add a new followerClock block in the profiler info. This is only set
when the follower could be a driver and it contains the clock info used
for following the driver, mostly the rate difference and delay.
Dump this info in pw-profiler -J
Make sure we always set the info in the clock, especially also when we
are following.
Keep a running average and variance of the error. Use this to
periodically update the DLL bandwidth. When the variance gets smaller,
we update the DLL more slowly to stay closer to the ideal rate.
This seems to improve the rate stability.
Use the helper instead of duplicating the same code.
Also add some helpers to parse a json array of uint32_t
Move some functions to convert between type name and id.
Add spa_json_begin_array/object to replace
spa_json_init+spa_json_begin_array/object
This function is better because it does not waste a useless spa_json
structure as an iterator. The relaxed versions also error out when the
container is mismatched because parsing a mismatched container is not
going to give any results anyway.
We have to unlink pcms when they are linked to a driver from
a different pcm.
When a playback and a capture pcm is linked and we start
the playback pcm and the capture pcm later this can leads
to a 'EPIPE' error on the capture device.
...
spa.alsa: hw:3,0c: snd_pcm_start: Broken pipe
...
While the spec allows for 1ppm changes, our rate matching logic applies
these changes quite often, which can be spammy on USB. I haven't seen
hosts mind this, but it seems like it might be a problem at some point.
Additionally, if we also have bind ctls enabled, every pitch update is
also a wakeup for ourselves (whether or not we're listening for the
pitch ctls, since the mixer fd does not distinguish between ctls, those
are filtered after we wake up).
The 10ppm threshold is empirically tested as being not "too noisy" (i.e.
when updates happen, I can see them scroll by with `amixer events`).
If necessary, we can make this configurable in the future.
Can be used to group ports together. Mostly because they are all from
the same stream and split into multiple ports by audioconvert/adapter.
Also useful for the alsa sequence to group client ports together.
Also interesting when pw-filter would be able to handle streams in the
future to find out what ports belong to what streams.
When bound_ctl info cannot be read this array elem info
is set to NULL in 'fetch_bind_ctl'. So when we iterate
the bound_ctl array we always have to check this.
In ACP mode, we might be accessing front:0 as the PCM, and using that
string to generate the ctl device name does not make sense. In
PulseAudio, we used the card index to generate a hw:X string, and we
replicate that here.
Fixes: https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/4028
This fixes an issue introduced in 771f71f622
where the quantum is forced and may break applications the specify their
own quantum.
Signed-off-by: Lukas Rusak <lorusak@gmail.com>
This patch fixes use case, when disable_tsched is set and
api.alsa.period-size is set to value different from default quantum size.
In a such configuration, threshold needs to be set to a final value
before snd_pcm_sw_params_set_avail_min is called to get IRQs with
right timing.
Avail minimum is calculated from a threshold set in the check_position_config.
The method returned different value for threshold right before playback
started and after the playback started. Therefore threshold used in
the snd_pcm_sw_params_set_avail_min was incorrect.
Force the check_position_config to use configured values when called
from spa_alsa_prepare as this method is called when starting new playback
and the state->period_frames and the state->rate are already known.
Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
Without this change, playback_ready or capture_ready was called
immediately after spa_alsa_start even tho start-delay was set.
Ready function was called with not precise "nsec" value, as "nsec"
plus latency should return time when the next buffer should be played
which wasn't true as start-delay was not included.
Now the playback is started immediately when the start_delay is set.
The alsa_do_wakeup_work is still called immediately but two things can
happened. Either start-delay is smaller then max_error and *_ready
function is called immediately, or start-delay is bigger then max_error
and state->next_time will be updated to correct value.
Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
Alsa needs to call handler soon enough to have headroom plus threshold
frames in the buffer and not only threshold left.
Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
Headroom are extra samples available in alsa buffer on top of a threshold.
Its use to prefill alsa buffer with silence before the playback starts
and later its use to calculate target number of a frames in the alsa buffer
when get_status is called. Target is calculated as headroom plus
threshold, which should be smaller then buffer size to make sense.
Signed-off-by: Martin Geier <martin.geier@streamunlimited.com>
Signed-off-by: Carlos Rafael Giani <crg7475@mailbox.org>
This does a couple of things: first, we implement revents demangling,
which seems to be required (although hw: devices work fine without it).
The second is to actually read the ctl events so we can tell when
elements we care about have changed, instead of reading everything and
trying to do a diff.
The latter is also required from a correctness perspective, as otherwise
the ctl might keep triggering wakeups while the fd is ready to be read.
Variable declarations after a label are not allowed,
and clang does not accept them. Fix the build failure
by removing the variable.
Fixes b3fbd0e607 ("alsa-pcm: add_bind_ctl_param: add support for array")
For matching kctl without the numid you need to specify interface,
device, subdevice, name and index. So the current implementation can
only match kctls on IFACE_PCM, device 0, subdevice 0 and index 0.
Instead of adding all these matching parameters this commit fetches all
kctls attached to the audio card and match on the first occurred kctl
with matching name.
This should be sufficient for audio cards with unique kctl names. When
non unique names are needed, more kctl matching parameters needs to be
added.
ALSA controls can only be opened on the card itself and will fail when
trying to open controls on the ALSA device. The device name we get may
or may not include the device suffix. If no suffix is present the
default device is 0 that's why currently it works on most audio cards.
But all other devices above 0 needs the suffix [1].
[1]
Device 0: hw:cardname
Device 0: hw:cardname,0
Device 1: hw:cardname,1
Device 2: hw:cardname,2
Device X: hw:cardname,X
We know in IRQ mode that any valid hi-res timestamp that the
driver privides will be before the wakeup event in pipewire.
This makes it so in IRQ mode we use better timestamping when possible,
which decreases jitter injected into the DLL, which in turn reduces
the amount of oscillations the resampler is exposed to.
We would timestamp within an unlikely block, which would introduce
additional jitter to current_time, which would have an impact on
the performance of the timer sensitive code.
Always reevaluate the rate matching even when we did not change the
follower state.
It is possible that we were a follower from some node with the same
clock and now become a follower of a node with a different clock. The
follower state doesn't change but we need to activate the rate matching
logic in that case.
Fixes rate matching in pro audio (playback) when capture and playback
are moved to another driver.
We can increase the MAX_LATENCY again if we increase the amount of
buffers when we are using a small buffer.
Normally we ask for 4 * quantum-limit as the buffer. This should be good
to use 1 buffer and quantum-limit as the quantum with enough headroom
to not run out of buffers.
If we are however using less buffer-frames we need to be careful and
allocate an extra buffer. Imagine using a buffer of 4096 frames, we can
support a quantum of up to 2048 frames if we use 2 buffers.
See #3744
Half of the buffersize is not enough to support as a max-quantum, we
need to divide by (4 * frame_scale) to allow some headroom and account
for the DSD scaling. We do the same calculation to suggest a buffer size
using the quantum-limit.
See #3744
As part of the setup for IRQ based scheduling, a period event
was installed. Not only is a timer based polling unecessary for
IRQ scheduling, depending on the state of the system, the timer
could fire far enough from the IRQ, causing alsa wakeup events
with no data in the ring buffer. Pipewire would identify these
events as an "early wakeup", adding an extra quantum of time
to the next_time estimate, skewing the clock and causing issues
with apps that depend on precise timing.
This reverts commit 49cdb468c2.
We should not do this, the nsec field should be relatable to the clock
monotonic time. If we use the estimated time, without actually using it
as a timer, we might end up with a wakeup time in the future compared to
the MONOTONIC clock time.
Instead, you can use the estimated current time simply by subtracting
the rate corrected duration from the next_nsec. This is really only
useful for some selected use cases (like in the JACK library).
This fixes some issues where in pro-audio mode, a client would try to
compare the current MONOTONIC time to nsec and find that it is in the
past.
This commit was done in an attempt to fix#3657 but it turned out the
real problem was something else.
Don't try to multiple the max_buffer_size with the frame scale or else
we might try to set a min_buffer_size larger than the max_buffer_size.
Instead, use the frame_scale only to scale the quantum_limit and then
clamp against the max_buffer size.
See #3000
We place the currently configured runtime settings in the properties but
clear the values when not negotiated. Don't do that but place the
defaults (when set) instead.
It doesn't make sense to hang these on the data loop, so let's have
these on the main loop instead. Also avoids a potential crash while
removing them (since removal happens on the main loop and the data loop
might be polling while we're doing the remove).
Don't use the current time as the nsec field in the graph clock because
it can jitter a lot. Instead, use the smoothed next_time, like we do
for timer based scheduling.
Since we track the current time against the rate converted ideal time,
lock on to the first timestamp when we reset the dll.
See #3657
Don't return the value of the last snd_pcm_resume() call because that
might be -ENOSYS when resume is not implemented for the card and then
the non-error (because we used drop/prepare later) propagates and
logs an error.
