Use the same sorting as ardour for midi events with the same timestamp
so that the order is:
Controller messages > Program Change > Note Off > Note On >
Note Pressure > Channel Pressure > Pitch Bend
Fixes#1868
If there is no valid channel map, assume MONO channels. A valid channel
map should be assigned at the source of the data, not here.
The problem is that when a device uses AUX channels, this will be fixed
up here with a surround setup, which is not right.
It is not enough for `buffer` to be alive in its current
scope because when execution enters that branch, `format`
will be set to `fmt`, which points inside `buffer`. And
since `format` is used outside that scope, `buffer` must
live longer.
This was detected by ASAN when Audacity was starting up.
==25007==ERROR: AddressSanitizer: stack-use-after-scope on address 0x7ffdbcfef560 at pc 0x7fe44ca95db3 bp 0x7ffdbcfeeda0 sp 0x7ffdbcfeed90
READ of size 4 at 0x7ffdbcfef560 thread T0
#0 0x7fe44ca95db2 in spa_pod_parser_pod ../spa/include/spa/pod/parser.h:67
#1 0x7fe44ca9a805 in spa_format_parse ../spa/include/spa/param/format-utils.h:44
#2 0x7fe44cad293a in port_set_format ../spa/plugins/audioconvert/audioconvert.c:1934
#3 0x7fe44cadad14 in impl_node_port_set_param ../spa/plugins/audioconvert/audioconvert.c:2038
#4 0x7fe44ca587e2 in configure_format ../spa/plugins/audioconvert/audioadapter.c:509
#5 0x7fe44ca60dff in negotiate_format ../spa/plugins/audioconvert/audioadapter.c:822
#6 0x7fe44ca62bbf in impl_node_send_command ../spa/plugins/audioconvert/audioadapter.c:846
#7 0x7fe45ea1c2f1 in node_update_state ../src/pipewire/impl-node.c:407
#8 0x7fe45ea5137e in pw_impl_node_set_state ../src/pipewire/impl-node.c:2251
#9 0x7fe45eb3355f in pw_work_queue_destroy ../src/pipewire/work-queue.c:142
#10 0x7fe45b2cd6f4 in source_event_func ../spa/plugins/support/loop.c:615
#11 0x7fe45b2c634f in loop_iterate ../spa/plugins/support/loop.c:452
#12 0x7fe45e9ebebc in spa_hook_list_clean ../spa/include/spa/utils/hook.h:395
#13 0x5561e03dc722 in main ../src/daemon/pipewire.c:131
#14 0x7fe45da3c28f (/usr/lib/libc.so.6+0x2328f)
#15 0x7fe45da3c349 in __libc_start_main (/usr/lib/libc.so.6+0x23349)
#16 0x5561e03db2a4 in _start ../sysdeps/x86_64/start.S:115
Address 0x7ffdbcfef560 is located in stack of thread T0 at offset 160 in frame
#0 0x7fe44ca56fa9 in configure_format ../spa/plugins/audioconvert/audioadapter.c:475
This frame has 4 object(s):
[32, 36) 'state' (line 493)
[48, 56) 'fmt' (line 494)
[80, 128) 'b' (line 492)
[160, 4256) 'buffer' (line 491) <== Memory access at offset 160 is inside this variable
When reading the timerfd gives an error, we should return right away
because the timeout did not happen.
If we change the timerfd timeout before reading it, we can get -EAGAIN.
Don't log an error in that case but wait for the new timeout.
This reverts commit 5bda4b6a57.
The reason is that when you use a null-audio-sink as a source, the
adapter will think it's a sink while wireplumber will try to use it as a
source.
There is no quick solution for this so revert this check for now.
We can only write from one thread to the ringbuffer so bypass the
ringbuffer when doing in-thread invoke. Only flush the current
items so that out-of-thread items don't get inserted.
Always append the item to the ringbuffer, even if we are invoking from
the thread itself. This ensure all items are always invoked in the
right order.
If we invoke from the thread, flush all items of the ringbuffer and
return.
Make sure to set the callback to NULL before invoking so that recursive
invoke doesn't call it again.
When while flushing the items we get a recursive invoke, detect this
with a counter and return immediately.
Mostly useful for when invoking from the thread itself so that the new
invoke item is executed before new items are added.
Imagine this case with module-loopback:
- data-loop goes into the capture process function
- mainloop invokes node remove of capture and waits
- data-loop invokes trigger -> node remove is first executed, mainloop
is woken up
- mainloop continues
- mainloop invokes remove of playback and waits
- data-loop continues flushing the ringbuffer -> playback remove is
executed, mainloop wakes up
- mainloop continues destroying items, frees playback
and capture streams
- data-loop finaly gets to flush the trigger and crashes because
streams are gone.
`Transform::Rot90` means the client should rotate 90 deg. clockwise,
which matches `SPA_META_TRANSFORMATION_90`, i.e. the buffer was
rotated 90 deg. anti-clockwise. The flipped cases should be correct
though.
Also add the source value to the debug print for easier future
debugging.
Fixes fa799aac86
First clear the started flag so that we ignore scheduling from the
follower. Then stop the follower and the converter.
This is the sequence we follow when deactivating a node, so do the
same here.
it is important that the node is not scheduled anymore when we clear
the format in suspend or else we might crash.
See #2877
And... we're back to 48Khz probing. Some devices fail to probe with
44.1KHz so when we need to choose between 2 bad things we choose to
do the right thing, which is probe in 48KHz.
Fixes#2857
libcamera can detect camera transforms/rotation, e.g. from the device
tree, and makes that information usable for clients via
`CameraConfiguration::transform`.
Advertise this information via the VideoTransform meta so Pipewire
clients can adjust their output accordingly.
Rotated cameras are common in mobile devices such as the Pinephone Pro,
which was used to test this feature.
When in passthrough mode and there is no converter, simply return
our own PortConfig parameter with our direction and passthrough mode.
Otherwise, use the PortConfig from the converter but filter out only the
PortConfig that matches our direction.
This fixes enumeration of PortConfig on the adapter.
When we try to read one of the events and there was an error, don't
signal the callback. If the error is something else than EAGAIN log
a warning.
Especially for timerfd, EAGAIN can happen when the timer changed
while polling. This can happen when running the profiler because it
polls and updates the timer from different threads.
Make a new variable to iterate the other ports so that we can use the
original port to emit notifications.
Fixes Latency and other params set on DSP ports.
When we are simply enumerating the params, just collect the info
but don't act on them, like clearing the format or recalculating
the latency.
This avoids some useless work when enumerating params.
Use the more complete configure_format function to clear the format
and buffers when the EnumFormat param changed.
When the follower updates EnumFormat, it probably wants to renegotiate
to a new format, so clear the current format so that we do that when
starting the next time.
EnumFormat should also not only return the current format in case we
are negotiated but it should return all possible formats.
See #2832
When we try to play data but the ringbuffer is full, we need to start
the device or else we will stay in this situation forever and stay
silent.
Fixes#2830
After the ports are reconfigured, we need to perform the setup again so
that buffers and processing can happen with the right settings.
This fixes an issue when autoswitching between A2DP and HFP with
bluetooth headsets when there is also a stereo capture device available.
The input stream of the browser is quickly reconfigured between stereo
and mono with only a Pause command in between, clearing the setup state
is enough to redo the setup when going back to Playing.
Fixes#2764
Find a free id from the available ids instead of just taking the current
number of devices, which would give assign the same id to multiple
cameras when: added 0, added 1, removed 0, added 1.
