In OpenEmbedded the PulseAudio recipe currently disables Valgrind
support by passing "ac_cv_header_valgrind_memcheck_h=no" to the
configure script (this was added to make it deterministic whether
Valgrdind support gets enabled or not). I'm converting the PulseAudio
recipe to use Meson, and I needed an option to disable Valgrind.
This patch adds the PA_COMMAND_SEND_OBJECT_MESSAGE command to protocol-native
so that clients can use the messaging feature introduced in the previous patch.
Sending messages can in effect replace the extension system for modules. The
approach is more flexible than the extension interface because a generic string
format is used to exchange information. Furthermore the messaging system can be
used for any object, not only for modules, and is easier to implement than
extensions.
Part-of: <https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/51>
The .include meta command already supports specifying a directory and
when including a directory, all files with the extension '.pa' in that
directory will be parsed in alphabetical order.
This feature can be used to add support for default.pa.d directory, so
that packages for other applications or users can just drop in a file
for configuration without changing the default.pa which is shipped.
We use the PA_DEFAULT_CONFIG_DIR for this, however, since meson quotes
this build variable, introduce an unquoted version for this purpose and
use it with .include.
Fixes: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/909
Signed-off-by: Sanchayan Maity <sanchayan@asymptotic.io>
With GStreamer 1.18, the old behaviour of storing the capture time in
DTS is gone (which is reasonable, since the semantics really don't
match). So instead, we get a capture timestamp when the buffer is being
pushed from udpsrc. This should eventually move into udpsrc, and the
timestamp should come from the cmsg instead of the clock.
We still fallback to the DTS if the meta isn't available, as the meta
might be dropped in older versions of rtpL16pay due to a bug.
Pipewire has started shipping copies of PulseAudio's ALSA card profiles.
It would be useful if both projects could share the same profiles and
this patch is a step toward that.
Prior to commits f899d5f466 and
f62a49b8cf, GNOME's sound settings
overwrote the routing for all entries in the stream-restore database
when selecting a device. Now we prevent that from happening (see the
aforementioned commits), but the old overwritten settings can still be in
the database after updating to PulseAudio 14.0, and they can cause
problems, as documented here:
https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/832
We can't distinguish between devices set by GNOME's sound settings
and devices set by the user, so this patch discards all old device
settings, even though that is going to cause PulseAudio to forget routing
settings for many users. This is less bad than keeping the incorrect
routing settings in the database, because it's difficult for users to
figure out how to fix the situation when e.g. speaker test tones go to
the internal speakers no matter what device is selected as the default,
whereas old manual configuration can be restored restored by doing the
manual configuration again. Also, it's probably more common to have at
some point changed the default device in GNOME's sound settings than it
is to have any manual per-stream routing settings.
This is disabled by default, because this causes data loss, but
distributions that use GNOME are recommended to enable this with
the --enable-stream-restore-clear-old-devices (Autotools) or
-Dstream-restore-clear-old-devices=true (Meson) build option.
Fixes: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/832
This adds a GStreamer-based RTP implementation to replace our own. The
original implementation is retained for cases where it is not possible
to include GStreamer as a dependency.
The idea with this is to be able to start supporting more advanced RTP
features such as RTCP, non-PCM audio, and potentially synchronised
playback.
Signed-off-by: Arun Raghavan <arun@arunraghavan.net>
Subdirectories add to the top-level cdata (specifically, the SIMD
detection happens in the pulsecore meson.build), so we were missing
HAVE_MMX/SSE2/NEON defines without this fix.
pulseaudio does not link against libbluetooth, as it's only talking to the
bluez daemon over dbus. So the build dependency on libbluetooth is overly
restrictive, as some embedded systems choose to ship without libbluetooth
but still have bluez daemon support.
This syncs the meson to the autotools configuration behavior by changing
the bluez option to a default on boolean.
For ease of maintaining both build systems, use the same version info
sequences as configure.ac. This should be simplified after Autotools has
been dropped.
- Rename "pulsedspdir" to the same "padsplibdir" that Autotools uses.
- Add a new option "pulsedsp-location" that is only used for padsp.in,
just like Autotools' --with-pulsedsp-location.
- Use 'set' instead of 'set_quoted' to avoid PULSEDSP_LOCATION getting
quoted twice.
Without this, meson on Solaris exits with:
src/daemon/meson.build:138:15: ERROR: Unknown variable "systemd_dep".
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
The original atomic implementation in pulseaudio based on
libatomic stated that the intent was to use full memory barriers.
According to [1], the load and store implementation based on
gcc builtins matches sequential consistent (i.e. full memory barrier)
load and store ordering only for x86.
I observed random crashes in client applications using memfd srbchannel
transport on an armv8-aarch64 platform (cortex-a57).
In all those crashes the first read on the pstream descriptor
(the size field) was wrong and looked like it contained old data.
I boiled the relevant parts of the srbchannel implementation down to
a simple test case and could observe random test failures.
So I figured that the atomic implementation was broken for armv8
with respect to cross-cpu memory access ordering consistency.
In order to come up with a minimal fix, I used the newer
__atomic_load_n/__atomic_store_n builtins from gcc.
With
aarch64-linux-gnu-gcc (Linaro GCC 7.3-2018.05) 7.3.1 20180425
they compile to
ldar and stlxr on arm64, which is correct according to [1] and [2].
The other atomic operations based on __sync builtins don't need
to be touched since they already are of the full memory barrier
variety.
[1] https://www.cl.cam.ac.uk/~pes20/cpp/cpp0xmappings.html
[2] <https://community.arm.com/developer/ip-products/processors
/b/processors-ip-blog/posts/armv8-a-architecture-2016-additions>
