pipewire-deb

Fork

Merge branch 'debian/master' into dyknon/master

Release pipewire version 0.3.85-1

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Bump debian/changelog

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Merge branch 'upstream/latest' into debian/master

Merge tag '0.3.85' into upstream/latest

0.3.85

bluez5: avoid compiler warning

Revert "gstpipewiresrc: break out of wait_started() also on STATE_UNCONNECTED"

This reverts commit 7465175ad0eea825928b47182b3fd0131fe2154d.

wait_started() is called before the stream is connecting and so
exits with an error immediately, which then makes the stream start
too early and block.

module-rt: Remove RLIMIT_RTTIME while trying rt policy

Trying rt scheduling policy while RLIMIT_RTTIME is set could result in
SIGXCPU if the limit is exceeded. Set an infinite limit temporarily
while checking if rt policy is allowed.

alsa: the default poll_descriptors is good now

Remove our custom poll_descriptors function because the default is
enough now that we update from revents.

alsa: update pollfd from poll_revents() as well

Some clients don't call _poll_descriptors() before entering the poll and
so don't get updated fds. This would result in a whole bunch of wakeups
when the eventfd was set but the new state is no longer active.

Part of the problem is also that check_active() returns the desired new
state and update_active() does check_active() + update. We could be
getting into a new desired inactive state without updating the eventfd.
Improve this by merging them together into update_active() which only
writes/reads the eventfd when something changes.

This also makes things less heavy because the eventfd is only
read/written when needed.

Fixes #3648

modules: add tunnel.may-pause option for pipe-tunnel

To override the default may-pause behaviour of streams.

modules: add pause support for pipe-tunnel

Streams are allowed to pause/resume when the pipe xruns.
Wait for IO_IN or IO_OUT when the ringbuffer xruns to resume playback or
capture.

Fixes #3197

alsa: return 0 from successful resume

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.

alsa: handle poll error

Recover after a poll error. This fixes resume after suspend.

Fixes #3646

alsa: remove err argument in alsa_recover()

We don't actually use it.

alsa: improve debug

jack: fix compilation with -UFASTPATH

bluez5: set BAP QoS RTN correctly

Set QoS RTN according to values supported by server values.

modules: make better media.name for RAOP sink

Use the destination in the media.name so that it becomes unique and the
volumes can be restored per destination.

Fixes #3801

bluez5: media-sink: set node.group for ISO group nodes

It's better to schedule nodes in the same ISO group together, since they
are supposed to produce synchronized output.

bluez5: iso-io: get proper group id for broadcast

bluez5: make iso_io ISO groups transport-specific

Different BT adapters have separate ISO groups, so we should schedule
them separately.

modules: Check for opus custom functions in addition to headers

alsa: reset ready_offset when we clear the queue

So that we start from the first sample of a new ready buffer.

jack: Add jack_port_t* <-> object helpers

Improve get_buffer debug.
Add some extra checks for the type before we try to use the object as a
port.
Set latency range to 0 when called with wrong port. Ardour seems to call
into us with ports from a previous jack_client..

jack: make 0 an invalid object type

So that we don't accidentally think uninitialized objects are ports.

spa: alsa: fix leaks in `get_data_path()`

Coverity rightfully complains that assigning to `path`
will lose its previous value leading to resource leaks.

gstpipewiresrc: break out of wait_started() also on STATE_UNCONNECTED

When the session manager sends an error to the client, it typically
also destroys the node after the error, which causes the stream to go
to STATE_UNCONNECTED via proxy_removed(). In that case, make sure
we exit the loop early, otherwise it will take 30 seconds to unblock
gst_element_set_state()

gst: avoid reporting error twice

First, make the error permanent by calling pw_stream_set_error()
and when this emits an error state again, report that to GStreamer.

Do the same in pipewiresink, which didn't even have the
pw_stream_set_error() call before, so the stream wasn't really going
into an error state at all.