This flag results in calls to (at least) isfinite() and isnan() becoming
skipped, and a constant false returned. This caused volume-test to fail
on Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=916504
Since PulseAudio deals with negative infinities with volume dB values,
this is not a problem only in volume-test. We shouldn't use -ffast-math
at all.
The recent change in ALSA upstream stripped -I$include/alsa path from
pkgconfig. We already fixed for this change in some places but still
the code for UCM was overlooked, and this resulted in the unresolved
symbols in alsa card module. Fix them as well.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The old minimum version was set in commit 57e3ccaf51 based on what the
commit author happened to have installed at the time. Russell Treleaven
now confirmed that Debian 8's gettext version, 0.19.3, works fine too,
or at least PulseAudio builds without errors. There might be room to
lower the required version even further, but that requires someone to
test older gettext versions.
This allows us to disable automatically updating build system files in
case things change. This is desirable in the common case, but not
necessarily for CI, where we want the ability to take a build directory
as an artifact from one stage to the next (i.e. into a fresh checkout).
It is helpful to improve reproducibility build [1] since
PA_SRCDIR/PA_BUILDDIR contains build path,
--disable-running-from-build-tree could drop these macros at
precompilation.
[1] https://reproducible-builds.org/
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
BlueZ 4 is no longer supported by BlueZ community for a long long time,
also by moving to BlueZ 5 it should make it even more clearer that
BlueZ 4 is no longer an option.
A new paprefs release is expected soon, and it will only support
GSettings. In order to have the default configuration work with the new
paprefs version, we need to enable GSettings by default.
If both module-gconf and module-gsettings are enabled when building
PulseAudio, both modules will be part of the default configuration. That
can cause trouble, because when the GConf database is migrated to
GSettings, the old configuration in GConf is not removed, so both
module-gconf and module-gsettings will try to load modules.
Generally it's not necessary to have both modules enabled even at build
time, so let's default to having only one of them enabled.
This also makes the module-gconf section conditional on HAVE_GCONF,
because if only gsettings support is built, the gconf section in the
configuration file would be redundant and confusing.
GConf is deprecated, and distributions are removing it. paprefs depends
on GConf, so in order to avoid paprefs getting removed as well, paprefs
has to be changed to use something else than GConf. GSettings is the
easiest alternative to migrate to, although it has the same problems
that GConf had: no support for system mode or networking.
This patch takes the non-GConf specific code from module-gconf and puts
it in stdin-util.[ch], which is then reused by module-gsettings.
module-gsettings is designed to be very similar to module-gconf.
Migration is expected to happen as follows: Distributions update
PulseAudio and paprefs at the same time, or first PulseAudio and then
paprefs. paprefs depends on module-gsettings, and module-gsettings
conflicts with module-gconf. Therefore module-gconf gets automatically
removed during the paprefs update. After the update an old PulseAudio is
likely to be running with module-gconf loaded. If the user tries to use
paprefs during this period, whatever the user does in paprefs won't have
any effect until PulseAudio is restarted (probably by a reboot or
relogin). This is not ideal, but will have to do.
When module-gsettings is loaded, it runs gsettings-data-convert
(implemented in a later patch). That will copy the settings from GConf
to GSettings. If gsettings-data-convert is not available (it's part of
GConf, so it may have already been uninstalled), then any previous
paprefs settings are lost.
glibc 2.27 is to be released soon, and it will provide memfd_create().
If glibc provides the function, we must not define it ourselves,
otherwise building fails due to conflict between the two implementations
of the same function.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=104733
Usually PulseAudio is built with a linker that supports the -z,now
option, and that option should have the same effect (i.e. the dynamic
linker resolves all symbols when the program is started) as re-execing
with the LD_BIND_NOW environment variable set, so usually the re-execing
is redundant.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=104789
The get_cpuid() function in cpu-x86.c was buggy on x86-64. When building
without optimizations, the homegrown assembly code overwrote the
beginning of the function argument list on the stack. That happened to
work fine on regular x86-64, but caused crashing with the x32 ABI.
At least GCC and clang provide cpuid.h, which has the __get_cpuid()
function that can be used instead of the homegrown assembly.
The PA_REG_* constants can be removed as well, because they're not used
any more.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=103656
This avoids the following autoconf warning:
configure.ac:89: warning: AC_COMPILE_IFELSE was called before AC_USE_SYSTEM_EXTENSIONS
../../lib/autoconf/specific.m4:368: AC_USE_SYSTEM_EXTENSIONS is expanded from...
configure.ac:89: the top level
The current build script hardcodes the $pkglibdir in the padsp command.
This works and is a reasonable default. However, distributions that
know where they install, can override this path and thus make padsp
work for any architecture that has the library installed by using the
following configure argument:
--with-pulsedsp-location='/usr/\\$$LIB/pulseaudio'
This works because ld.so considers $LIB a variable that will expand to
several location paths, depending on the architecture of the binary
being executed.
In debian, for example, this would work for libpulsedsp.so installed in
/usr/lib/x86_64-linux-gnu/ for amd64 and /usr/lib/i386-linux-gnu/ for
i386, with a single padsp command.
Any compiler flags should be set before asking the compiler to check for
thread-local storage with AX_TLS, since compiler flags (in this case
-mmacosx-version-min=10.5) can influence the outcome of that check.
The module doesn't build any more[1], and when starting to investigate
the build failure, I asked the module author if he'd know something
about the breakage. He said that he didn't know about backward
compatibility problems with libxen, but more importantly, he said that
the module probably doesn't have any users[2]. It doesn't make sense to
keep maintaining a module that doesn't have users, so let's drop it.
[1] https://bugs.freedesktop.org/show_bug.cgi?id=98793
[2] https://lists.freedesktop.org/archives/pulseaudio-discuss/2016-November/027172.html
Although such 9.0 clients support memfd transport, they have an
iochannel bug that would break memfd audio if they're run in 32
bit mode over a 64-bit kernel. Influence them to use the POSIX
shared memory model instead.
Also bump the protocol version to exclusively mark such v9.0
libraries. Check commit 451d1d6762 for further details.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=97769
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
tests/core-util-test.c uses ck_assert_int_lt() which was introduced
in check 0.9.10
make this dependency (with --enable-tests) explicit in configure.ac
Signed-off-by: Peter Meerwald-Stadler <pmeerw@pmeerw.net>
json-c has a symbol clash (json_object_get_type) with json-glib (which
at least a number of our GNOME clients use). This patch moves to our own
JSON parser so that we can avoid this kind of situation altogether.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=95135
Signed-off-by: Arun Raghavan <arun@arunraghavan.net>
Now that all layers in the stack support memfd blocks, add memfd
support for the daemon's global core mempool. Also introduce
"enable-memfd=" daemon argument and configuration option.
For now, memfd support is an opt-in feature to be activated only
when daemon's enable-memfd= is set to yes.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Per glibc feature_test_macros(7), setting compiler flags to
-std=c11 (or any c* variant like c99) enforces strict ANSI
mode.
Enforcing strict ANSI makes all declarations under _GNU_SOURCE
unavailable. This leads to build warnings in the form of:
warning: implicit declaration of function ‘syscall’
Thus replace -std=c11 with -std=gnu11
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Memfd is a simple memory sharing mechanism, added by the systemd/kdbus
developers, to share pages between processes in an anonymous, no global
registry needed, no mount-point required, relatively secure, manner.
This patch introduces the necessary building blocks for using memfd
shared memory transfers in PulseAudio.
Memfd support shall also help us in laying out the necessary (but not
yet sufficient) groundwork for application sandboxing, protecting PA
from its clients, and protecting clients data from each other.
We plan to exclusively use memfds, instead of POSIX SHM, on the way
forward.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
This is needed for building with anonymous unions. A bunch of calls to
fail() that used to mysteriously work need fixing -- fail() is a macro
that takes a printf-style message as an argument. Not passing this
somehow worked with the previous compiler flags, but breaks with
-std=c11.
There is no way to check CPU type in a portable way across ABIs.
Assume if pointers are 64-bit that CPU is capable to perform fast
64-bit operations. Add an extra check to handle x32-ABI.
PulseAudio by default builds with -Wundef. If we add -Werror=undef this
missing define is fatal. By default build log is full of entries like:
In file included from ./pulsecore/core.h:47:0,
from ./pulsecore/module.h:31,
from ./pulsecore/sink-input.h:31,
from pulsecore/sound-file-stream.c:36:
./pulsecore/sample-util.h: In function 'pa_mult_s16_volume':
./pulsecore/sample-util.h:58:5: warning: "__WORDSIZE" is not defined [-Wundef]
#if __WORDSIZE == 64 || ((ULONG_MAX) > (UINT_MAX))
^
(NetBSD-7.99.21 with default GCC 4.8.5)
This change fixes build issues on NetBSD.
This also address a bug reported by Shawn Walker from Oracle (possibly Solaris):
Bug 90880 - builds can fail due to non-portable glibc-specific internal macro usage
On some systems (at least Arch) DATADIRNAME is not defined. This
caused PULSE_LOCALEDIR to point to a wrong directory. This seemed like
an issue introduced in 7.0, but probably something else was updated in
Arch at the same time, causing DATADIRNAME to become undefined,
because there were no changes between 6.0 and 7.0 that could have
caused this.
After noticing that localedir is a standard variable, my first idea
was to use pulselocaledir='${localedir}' in configure.ac, but Jan
Steffens pointed out that it causes the final PULSE_LOCALEDIR to
become "${prefix}/share/locale", that is, the variables weren't fully
expanded. I then found a FAQ item in Autoconf's manual[1], which
recommends not to define any absolute installation directories in
configure.ac, because the installation directories should be possible
to change when running make. The recommended solution is to define the
constant in AM_CPPFLAGS instead, so that's what this patch does.
[1] https://www.gnu.org/software/autoconf/manual/autoconf-2.69/html_node/Defining-Directories.html
If bashcompletiondir was empty, the check didn't catch that. As
a result, the symlinks that were supposed to be generated in the
completion directory were created in the root directory.
The nodelete flag indicates that we don't want our libraries to be
unloaded. It's only relevant on libraries, so let's not use it for
executables. Trying to use it on executables breaks things on some
platforms.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=90878
