It was possible that auto-connect-thread or GUI thread called
DelayLine::set_delay without acquiring the process-lock.
Changing the delay while processing is not supported and
can cause glitches.
backtrace:
```
DelayLine::set_delay
Send::update_delaylines
Send::set_delay_out
InternalReturn::set_playback_offset
Route::update_signal_latency
Session::update_route_latency
Session::update_latency_compensation
Session::auto_connect_thread_run
```
"While 'atomic' has a volatile qualifier, this is a historical
artifact and the pointer passed to it should not be volatile."
Furthermore "It is very important that all accesses to a
particular integer or pointer be performed using only this API"
(from https://developer.gnome.org/glib/2.68/glib-Atomic-Operations.html)
Hence initialization of atomic variables is changed to also use
this API, instead of directly initializing the value.
This also fixes a few cases where atomic variables were
accessed directly.
see also libs/pbd/pbd/g_atomic_compat.h
This slowly un/mutes the direct monitor signal when listen state
changes.
Listen sends already fade in as well. However un-soloing still
hard-removes the listen send path.
When a route's output is connected to multiple busses, re-establishing
port-connections on session-load calls Route::output_change_handler
multiple times.
Implicit solo counts are saved and restored by the route's SoloControl,
however the output_change_handler invalidated it on session load.
Since connections are restored iteratively, counting soloed downstream
routes will yield an incorrect result until all connections are
established.
This fixes the issue that implicit solo is lost on session-reload
in case a route has multiple connections with different up/downstream
solo states.
* Include main-out channel-count when bouncing the output
* Do not de-activate disk-reader and mains-out when freezing a track
* Ignore meter when counting channels
This is only relevant when automating a plugin-bypass, due to
ActiveChanged the following could happen from the realtime thread:
#2 ARDOUR::Session::update_latency_compensation (false, false)
#3 PBD::Signal0<void, PBD::OptionalLastValue<void> >
#4 ARDOUR::AutomationControl::actually_set_value
#5 ARDOUR::PluginInsert::PluginControl::actually_set_value
#6 ARDOUR::PluginInsert::connect_and_run
#7 ARDOUR::PluginInsert::automate_and_run
In this specific case the update_latency_compensation()
is called with process-lock held but the caller lied:
called_from_backend=false
When a delayline update is needed, this can lead to a deadlock,
since the process-lock is acquired again.
Simply postponing the update is preferable since this also
postpones the change when actively recording.
PS. This may need special casing for freewheel export.
IO::latency iterates over the port-set retrieving the
private_latency_range of each port. The value can cached
since it can only change when connections or latency changes.
see also 40eefeddd6
Previously when the master-bus had more outputs than inputs,
Ardour crashed when the monitor-section was set up.
Removing a master-bus output port calls
Route::output_change_handler (master-bus)
-> Session::reset_monitor_section
which first removes the corresponding monitor-section input,
then output port. The latter triggers
ARDOUR::Route::output_change_handler (monitor-bus).
All with the process-lock held, so at this point in time Ardour
has removed the port-reference but the port still exists in the
backend.
Now the monitor-bus processors are re-configured and
the channel-count is updated. The port that was just removed
and triggered the ::output_change_handler() callback is
re-created.
unable to create port 'Monitor/audio_out 2': failed constructor
This fix changes the monitor-section to use strict-i/o (for plugins)
and also use master-bus output (not input) when configuring
processors.
Since fan-out happens (interactively) in the GUI thread,
we need to keep track of this, in order to prevent later
auto-connection of the track to the master-bus.
When removing a plugin-insert that has a sidechain on a bus
that is implicitly soloed the following happens:
* Route::remove_processor() takes a WriterLock(_processor_lock)
* The sidechain input is disconnected
* Route::sidechain_change_handler() is called which calls
* Route::input_change_handler()
* Since the route is implicitly soloed, propagation is attempted
which calls
* Route::direct_feeds_according_to_reality()
which takes ReaderLock(_processor_lock)
Recursive locks, reader-lock after writer-lock don't cause
a deadlock, releasing the reader-lock effectively also
releases the writer-lock...
It is possible that Route::monitoring_state() returns
(MonitoringDisk | MonitorSilence)
This lead to various cases where there were is a direct comparison
(ms == MonitoringDisk). DiskReader::run tests for MonitoringDisk
to check if the buffer needs to be zeroed while locating.
Likewise Route::process_output_buffers() also explicitly tests
for both MonitoringDisk and MonitoringDisk.
The issue was likely introduced in fbe8075117 (although
it may have been possible in earlier version when using hardware
monitoring as well).
Previously add_remove_sidechain() released the process-lock
(to create the I/O ports), while keeping a processor writer-lock.
Meanwhile the auto-connect thread may be woken up to perform
a latency update. This thread takes the process-lock and
then stalls, waiting for a processor reader-lock.
Then add_remove_sidechain() continues and tries to re-acquire
the process-lock.
This partially reverts 208c781248
in order to fix monitoring when using punch-in/out.
This also allows to revert to Ardour 5 style MIDI exclusive
Input or Disk monitoring when not using layered-recording.
The GUI uses ::axis_view_by_control() if a controllable when
the CoreSelection contains a controllable.
CoreSelection::get_stripables() saves and looks up controllable
by PBD::ID. Panorama automation controls were previously
not found, since they are not directly owned by the route.
So far internal-sends were explicitly special-cased, and implicit
solo-state only was only forwarded for direct up/down-stream
connections. ...and nobody can remember why :)
This restriction is removed, so far mainly with commenting-out
code, as experiment.
Internal sends are not only more convenient, but also required
in cases where direct connections result in ambiguous latency.
This fixed a false-positive "ambiguous latency" warning for
MIDI busses:
MIDI track -midi-> MIDI Bus w/instrument -audio|midi-> master
The master-bus only has audio inputs, so the MIDI Bus'
MIDI out is left unconnected. Its latency does not matter,
it can float freely.
Previously it was assumed to be zero. So the MIDI Bus' input
playback latency range was [0, master-bus playback-latency].
Various backends have different strategies for updating latencies,
and different thread semantics for latency update callbacks.
With jack2 it is possible that processing happens concurrently
while port latencies are being changed.
Ardour internal backends emit the latency callback from the
main process thread, serializing latency changes with processing.
Various user actions can also trigger latency changes. e.g.
with a stopped engine it is possible to re-order plugins.
IO::latency iterates over the port-set retrieving the
private_latency_range of each port. Since it can only change
when connections and latency changes, we can instead cache the value.
This is also in preparation to allow the auditioner to override it.
* Signals use camelcase.
* use a static signal, independent of the route
Previously the signal was handled by RouteUI, which
caused various issues:
* the RouteUI may not yet be available
* There may be many RouteUIs for a single Route
(mixer, editor-mixer, meter-bridge strips)
* Remove unused direct calls into plugin
* Assume empty model to mean plugin-provided MIDNAM (!)
The route owned Instrument-Info is the central access point used
by the GUI for MIDI name lookups.
At this point in time, custom settings are saved/restored by the
GUI (MidiTimeAxisView). InstrumentInfo provides a volatile store
for MIDNAM mode and model.
This fixes another assert(), caused by configuring processors,
before set_processor_state() was called.
Route::configure_processors() will be called later.
---
#3 0x00007ffff2472102 in __GI___assert_fail at assert.c:101
#4 0x00007ffff7a8ca1f in ARDOUR::Route::setup_invisible_processors() at ../libs/ardour/route.cc:5013
#5 0x00007ffff7a7a665 in ARDOUR::Route::configure_processors_unlocked at ../libs/ardour/route.cc:1870
#6 0x00007ffff7a79377 in ARDOUR::Route::configure_processors at ../libs/ardour/route.cc:1719
#7 0x00007ffff7a902c0 in ARDOUR::Route::set_disk_io_point at ../libs/ardour/route.cc:6041
#8 0x00007ffff7a7ea0a in ARDOUR::Route::set_state at ../libs/ardour/route.cc:2679
Previously set_state() -> set_meter_point() acquired the
process-lock to change meter-position, usually causing x-runs
when setting route-state.
This also fixes an issue introduced in fd414ec158. After
populating the processor list, force setting the meter-position
looks up the output streams of the processor before the meter.
However the processors are not configured. That will only happen
later from Session::post_engine_init().
---
#3 0x00007ff07b7d4102 in __GI___assert_fail at assert.c:101
#4 0x00007ff080d3224a in ARDOUR::PluginInsert::output_streams() const at ../libs/ardour/plugin_insert.cc:289
#5 0x00007ff080de8c30 in ARDOUR::Route::set_meter_point_unlocked() at ../libs/ardour/route.cc:4106
#6 0x00007ff080de8699 in ARDOUR::Route::set_meter_point(ARDOUR::MeterPoint, bool) at ../libs/ardour/route.cc:4037
#7 0x00007ff080ddfad3 in ARDOUR::Route::set_state(XMLNode const&, int) at ../libs/ardour/route.cc:269
This fixes the following issue:
On the master channel insert the waveform scope before the fader.
Then set the meter position to custom and move the meter to the
very beginning of the chain.
Before this change, when set_meter_point() was called the
processor list only contained the Fader (_amp) and no other
processor. _main_outs was not yet present in the list, and
Route::maybe_note_meter_position() triggered an
and assert(_processor_after_last_custom_meter.lock());
See also d0dca7daf0
This handles some special cases where a plugin is added after
the disk-writer but before the disk-reader.
The plugin may add/remove ports (e.g. an instrument: MIDI to audio,
or some stereo to mono processors). However we need to ensure
that any data that is recorded will be played back.
This is a new take replacing b2bc934e2.
[Re]name sidechain port directly when a plugin is added.
Usually plugins are constructed without an Route (owner is unset).
PluginInsert c'tor may already create a side-chain port, at the
time of construction the sidechain port will be created using
the port-name "toBeRenamed".
Previously the plugin had to be added to a route using "add_processor",
in order to set a unique name, before a new plugin with sidechain
could be constructed.
ProcessorBox::use_plugins() did that in the correct sequence,
however there may have been cases where this didn't work, and
Route::add_processors() was called directly..
This builds on top of 51d2bb:
* v6 routes templates/states have a version per <Route>
* older route-states are assumed to be from ardour-5
Stateful::loading_state_version 3002,
unless specified otherwise
Currently using Ardour-5 route templates (state version "3002")
with Ardour6 fails. As opposed to session-templates, Route
templates were not versioned.
This ensures future compatibility (and may allow to interpret
unversioned templates as "3002")
* update AFL position when preference changes
* "after post-fader processors (before pan)" is before
the main-out (not at the end).
* Fix "immediately post-fader":
The amp, when added was the last element. ++after_end then
made the iterator point to .end()
This likely worked in the past when the monitor send was added
immediately after adding the fader/amp before any other processors.
While loading a session XML state, set_state must use
`Stateful::loading_state_version`.
When later copying processor state,
`Stateful::current_state_version` is correct.
This also solves bi-stable automation for plugins where latency
can change due to automation. e.g.
cycle 1: run (t): automation (t) = on: -> increase latency
cycle 2: run (t-latency): automation (t-latency) = off -> decrease latency
repeat.
Paul Davis
Robin Gareus
Len Ovens
Ben Loftis