Add the buffer_get_free pcm_ops function to reduce the effective
playback buffer size. All intermediate audio playback buffers
become temporary buffers.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-11-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Commit ff095e5231 "audio: api for mixeng code free backends"
introduced another FIFO for the audio subsystem with exactly the
same size as the mixing-engine FIFO. Most audio backends use
this generic FIFO. The generic FIFO used together with the
mixing-engine FIFO doubles the audio FIFO size, because that's
just two independent FIFOs connected together in series.
For audio playback this nearly doubles the playback latency.
This patch restores the effective mixing-engine playback buffer
size to a pre v4.2.0 size by only accepting the amount of
samples for the mixing-engine queue which the downstream queue
accepts.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Message-Id: <20220301191311.26695-10-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This reverts commit cbaf25d1f5.
Since previous commit every audio backend has a pcm_ops function
table. It's no longer necessary to test if the table is available.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-9-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add a pcm_ops function table for the capture backend. This avoids
additional code in the next patches to test if the pcm_ops table
is available.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-8-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Change the code to copy the playback stream in sequential order.
The advantage can be seen in the next patches where the stream
copy operation effectively becomes a write through operation.
The following diagram shows the average buffer fill level and
the stream copy sequence. ### represents a timer_period sized
chunk. The rest of the buffer sizes are not to scale.
With current code:
|--------| |#####111| |---#####|
sw->buf mix_buf backend buffer
1. clip
|--------| |---#####| |111##222|
sw->buf mix_buf backend buffer
2. write to audio device
333 -> |--------| |---#####| |---111##| -> 222
sw->buf mix_buf backend buffer
3a. sw device write
|-----333| |---#####| |---111##|
sw->buf mix_buf backend buffer
3b. resample and mix
|--------| |333#####| |---111##|
sw->buf mix_buf backend buffer
With this patch:
111 -> |--------| |---#####| |---#####|
sw->buf mix_buf backend buffer
1a: sw device write
|-----111| |---#####| |---#####|
sw->buf mix_buf backend buffer
1b. resample and mix
|--------| |111##222| |---#####|
sw->buf mix_buf backend buffer
2. clip
|--------| |---111##| |222##333|
sw->buf mix_buf backend buffer
3. write to audio device
|--------| |---111##| |---222##| -> 333
sw->buf mix_buf backend buffer
The effective total playback buffer size is reduced by
timer_period.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-7-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The next patch reduces the effective qemu playback buffer size
by timer-period. Increase the number of jack audio buffers by
one to preserve the total effective buffer size. The size of one
jack audio buffer is 512 samples. With audio defaults that's
512 samples / 44100 samples/s = 11.6 ms and only slightly larger
than the timer-period of 10 ms.
The larger jack audio buffer increases audio dropout safety,
because the high priority jack-audio worker threads can provide
audio data for a longer period of time as with a smaller buffer
and more audio data in the mixing engine buffer that they can't
access.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Christian Schoenebeck <qemu_oss@crudebyte.com>
Message-Id: <20220301191311.26695-6-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This is a patch to improve the pulseaudio playback experience.
Asking pulseaudio for a playback latency of 15ms is quite
demanding. Increase this to 46ms. The total playback latency
now is 31ms larger. One of the next patches will reduce the
total playback latency again by more than 46ms.
Here is a quote from the PulseAudio Latency Control
documentation: 'For the sake of (...) drop-out safety always
make sure to pick the highest latency possible that fulfills
your needs.'
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-5-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Simplify code by inlining function audio_pcm_sw_get_rpos_in()
at the only call site and remove the duplicated audio_bug()
test.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-4-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add a function audio_pcm_hw_conv_in() similar to the existing
counterpart function audio_pcm_hw_clip_out(). This function reduces
the number of calls to the pcm_ops functions get_buffer_in() and
put_buffer_in(). That's one less call to get_buffer_in() and
put_buffer_in() every time the conv_buffer wraps around.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-3-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Move the function audio_pcm_hw_clip_out() into the correct
section 'Hard voice (playback)'.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20220301191311.26695-2-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Replace open-coded buffer arithmetic with the new function
audio_ring_posb(). That's the position in backward direction
of a given point at a given distance.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Message-Id: <20220301191311.26695-1-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Audio recordings with the DirectSound backend don't sound right.
A look a the Microsoft online documentation tells us why.
From the DirectSound Programming Guide, Capture Buffer Information:
'You can safely copy data from the buffer only up to the read
cursor.'
Change the code to read up to the read cursor instead of the
capture cursor.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20211226154017.6067-2-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
On Windows the jack_set_thread_creator() function and on MacOS the
pthread_setname_np() function with a thread pointer paramater is
not available. Use #ifdefs to remove the jack_set_thread_creator()
function call and the qjack_thread_creator() function in both
cases.
The qjack_thread_creator() function just sets the name of the
created thread for debugging purposes and isn't really necessary.
From the jack_set_thread_creator() documentation:
(...)
No normal application/client should consider calling this. (...)
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/785
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Christian Schoenebeck <qemu_oss@crudebyte.com>
Message-Id: <20211226154017.6067-1-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add a new -audio backend that accepts D-Bus clients/listeners to handle
playback & recording, to be exported via the -display dbus.
Example usage:
-audiodev dbus,in.mixing-engine=off,out.mixing-engine=off,id=dbus
-display dbus,audiodev=dbus
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
This brings a change that makes audio drivers more similar to all
other modules. All drivers are built by default, while
--audio-drv-list only governs the default choice of the audio driver.
Meson options are added to disable the drivers, and the next patches
will fix the help messages and command line options, and especially
make the non-default drivers available via -audiodev.
Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130630.632028-4-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Ever since winwaveaudio was removed in 2015, CONFIG_AUDIO_WIN_INT
is only set if dsound is in use, so use CONFIG_AUDIO_DSOUND directly.
Cc: Gerd Hoffman <kraxel@redhat.com>
Cc: Volker Rümelin <vr_qemu@t-online.de>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20211007130630.632028-3-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The audio migration vmstate is empty, and always has been; we can't
just remove it though because an old qemu might send it us.
Changes with -audiodev now mean it's sometimes created when it didn't
used to be, and can confuse migration to old qemu.
Change it so that vmstate_audio is never sent; if it's received it
should still be accepted, and old qemu's shouldn't be too upset if it's
missing.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Tested-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20210809170956.78536-1-dgilbert@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Jose R. Ziviani <jziviani@suse.de>
Message-Id: <20210624103836.2382472-9-kraxel@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
On macOS 11.3.1, Core Audio calls AudioDeviceIOProc after calling an
internal function named HALB_Mutex::Lock(), which locks a mutex in
HALB_IOThread::Entry(void*). HALB_Mutex::Lock() is also called in
AudioObjectGetPropertyData, which is called by coreaudio driver.
Therefore, a deadlock will occur if coreaudio driver calls
AudioObjectGetPropertyData while holding a lock for a mutex and tries
to lock the same mutex in AudioDeviceIOProc.
audioDeviceIOProc, which implements AudioDeviceIOProc in coreaudio
driver, requires an exclusive access for the device configuration and
the buffer. Fortunately, a mutex is necessary only for the buffer in
audioDeviceIOProc because a change for the device configuration occurs
only before setting up AudioDeviceIOProc or after stopping the playback
with AudioDeviceStop.
With this change, the mutex owned by the driver will only be used for
the buffer, and the device configuration change will be protected with
the implicit iothread mutex.
Signed-off-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Message-id: 20210622201740.38005-1-akihiko.odaki@gmail.com
Message-Id: <20210622201740.38005-1-akihiko.odaki@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Before commit 7d6948cd98, it was coded to
retrieve the initial output stream format settings, modify the frame
rate, and set again. However, I removed a frame rate modification code by
mistake in the commit. It also assumes the initial output stream format
is consistent with what QEMU expects, but that expectation is not in the
code, which makes it harder to understand and will lead to breakage if
the initial settings change.
This change explicitly sets all of the output stream settings to solve
these problems.
Signed-off-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Message-Id: <20210616141721.54091-1-akihiko.odaki@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Currently with jackaudio client name and qemu guest name unset,
the JACK client names are out-(NULL) and in-(NULL). These names
are user visible in the patch bay. Replace the function call to
qemu_get_vm_name() with a call to audio_application_name() which
replaces NULL with "qemu" to have more descriptive names.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20210517194604.2545-4-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Move the code to generate the pa_context_new() application name
argument to a function in audio/audio.c. The new function
audio_application_name() will also be used in the jackaudio
backend.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20210517194604.2545-3-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
In current code there are no calls to pa_stream_get_latency()
or pa_stream_get_time() to receive latency or time information.
Remove the flags PA_STREAM_INTERPOLATE_TIMING and
PA_STREAM_AUTO_TIMING_UPDATE which instruct PulseAudio to
calculate this information in regular intervals.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20210517194604.2545-2-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Merge the #ifdef DEBUG code with the if statement a few lines
above to avoid bit rot.
Suggested-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20210517194604.2545-1-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Commit e50caf4a5c ("tracing: convert documentation to rST")
converted docs/devel/tracing.txt to docs/devel/tracing.rst.
We still have several references to the old file, so let's fix them
with the following command:
sed -i s/tracing.txt/tracing.rst/ $(git grep -l docs/devel/tracing.txt)
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210517151702.109066-2-sgarzare@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
An output device change can occur when plugging or unplugging an
earphone.
Signed-off-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Message-Id: <20210311151512.22096-3-akihiko.odaki@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Mac OS X 10.6 was released in 2009.
Signed-off-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20210311151512.22096-1-akihiko.odaki@gmail.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The 'running' argument from VMChangeStateHandler does not require
other value than 0 / 1. Make it a plain boolean.
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20210111152020.1422021-3-philmd@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Delete spaces between function name and open parenthesis'('
Signed-off-by: Zhang Han <zhanghan64@huawei.com>
Message-id: 20210115012431.79533-1-zhanghan64@huawei.com
Message-Id: <20210115012431.79533-8-zhanghan64@huawei.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Fix the line width of code.
Signed-off-by: Zhang Han <zhanghan64@huawei.com>
Message-id: 20210115012431.79533-1-zhanghan64@huawei.com
Message-Id: <20210115012431.79533-5-zhanghan64@huawei.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Fix problems about braces:
-braces are necessary for all arms of if/for/while statements
-else should follow close brace '}'
Signed-off-by: Zhang Han <zhanghan64@huawei.com>
Message-id: 20210115012431.79533-1-zhanghan64@huawei.com
Message-Id: <20210115012431.79533-2-zhanghan64@huawei.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
There is a mismatch between message and used argument. Change
the argument from frequency to format.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-23-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Rename dsound_open() to dsound_set_cooperative_level(). The
only task of that function is to set the cooperative level for
DirectSound.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-21-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
GetForegroundWindow() doesn't necessarily return the own window
handle. It just returns a handle to the currently active window
and can even return NULL. At the time dsound_open() gets called
the active window is most likely the shell window and not the
QEMU window.
Replace GetForegroundWindow() with GetDesktopWindow() which
always returns a valid window handle, and at the same time
replace the DirectSound buffer flag DSBCAPS_STICKYFOCUS with
DSBCAPS_GLOBALFOCUS where Windows only expects a valid window
handle for DirectSound function SetCooperativeLevel(). The
Microsoft online docs for IDirectSound::SetCooperativeLevel
recommend this in the remarks.
This fixes a bug where you can't hear sound from the guest.
To reproduce start qemu with -machine pcspk-audiodev=audio0
-device intel-hda -device hda-duplex,audiodev=audio0
-audiodev dsound,id=audio0,out.mixing-engine=off
from a shell and start audio playback with the hda device in the
guest. The guest will be silent. To hear guest audio you have to
activate the shell window once.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-20-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Tell PulseAudio to send recorded audio data in smaller chunks
than timer_period, so there's a good chance that qemu can read
recorded audio data every time it looks for new data.
PulseAudio tries to send buffer updates at a fragsize / 2 rate.
With fragsize = timer_period / 2 * 3 the update rate is 75% of
timer_period. The lower limit for the recording buffer size
maxlength is fragsize * 2.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-19-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Currently with the playback buffer attribute minreq = -1 and flag
PA_STREAM_EARLY_REQUESTS PulseAudio uses minreq = tlength / 4.
To improve audio playback with larger PulseAudio server side
buffers, limit minreq to a maximum of 75% of audio timer_rate.
That way there is a good chance qemu receives a stream buffer
size update before it tries to write data to the playback stream.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-18-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The audio buffer size in audio/paaudio.c is typically larger
than expected. Just comment the bugs in qpa_init_in() and
qpa_init_out() for now. Fixing these bugs may break glitch free
audio playback with fine tuned user audio settings.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-17-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Don't call pa_stream_writable_size() in qpa_get_buffer_out()
before the playback stream is ready. This prevents a lot of the
following pulseaudio error messages.
pulseaudio: pa_stream_writable_size failed
pulseaudio: Reason: Bad state
To reproduce start qemu with
-parallel none -device gus,audiodev=audio0 -audiodev pa,id=audio0
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-15-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Don't call pa_stream_writable_size() in qpa_write() before the
playback stream is ready. This prevents a lot of the following
pulseaudio error messages.
pulseaudio: pa_stream_writable_size failed
pulseaudio: Reason: Bad state
To reproduce start qemu with
-parallel none -device gus,audiodev=audio0
-audiodev pa,id=audio0,out.mixing-engine=off
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-14-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The pulseaudio backend currently converts, clips and copies audio
playback samples in the mixing-engine sample buffer multiple
times.
In qpa_get_buffer_out() the function pa_stream_begin_write()
returns a rather large buffer and this allows audio_pcm_hw_run_out()
in audio/audio.c to copy all samples in the mixing-engine buffer
to the pulse audio buffer. Immediately after copying, qpa_write()
notices with a call to pa_stream_writable_size() that pulse audio
only needs a smaller part of the copied samples and ignores the
rest. This copy and ignore process happens several times for each
audio sample.
To fix this behaviour, call pa_stream_writable_size() in
qpa_get_buffer_out() to limit the number of samples
audio_pcm_hw_run_out() will convert. With this change the
pulseaudio pcm_ops functions put_buffer_out and write are no
longer identical and a separate qpa_put_buffer_out is needed.
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 9315afe5-5958-c0b4-ea1e-14769511a9d5@t-online.de
Message-Id: <20210110100239.27588-13-vr_qemu@t-online.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
