Now that we have the necessary infrastructure to memexport and
mempimport a memfd memblock, extend that support higher up in the
chain with pstreams.
A PA endpoint can now _transparently_ send a memfd memblock to the
other end by simply calling pa_pstream_send_memblock() – provided
the block's memfd pool was earlier registered with the pstream.
If the pipe does not support memfd transfers, we fall back to
sending the block's full data instead of just its reference.
** Further details:
A single pstream connection usually transfers blocks from multiple
pools including the server's srbchannel mempool, the client's
audio data mempool, and the server's global core mempool.
If these mempools are memfd-backed, we now require registering
them with the pstream before sending any blocks they cover. This
is done to minimize fd passing overhead and avoid fd leaks.
Moreover, to support all these pools without hard-coding their
number or nature in the Pulse communication protocol itself, a new
REGISTER_MEMFD_SHMID command is introduced. That command can be
sent _anytime_ during the pstream's lifetime and is used for
creating on demand SHM ID to memfd mappings.
Suggested-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Color global mempools with a special mark. This special marking
is needed for handling memfd-backed pools.
To avoid fd leaks, memfd pools are registered with the connection
pstream to create an ID<->memfd mapping on both PA endpoints.
Such memory regions are then always referenced by their IDs and
never by their fds, and so their fds can be safely closed later.
Unfortunately this scheme cannot work with global pools since the
registration ID<->memfd mechanism needs to happen for each newly
connected client, and thus the need for a more special handling.
That is, for the pool's fd to be always open :-(
Almost all mempools are now created on a per-client basis. The
only exception is the pa_core's mempool which is still shared
between all clients of the system.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Soon we're going to have three types of memory pools: POSIX shm_open()
pools, memfd memfd_create() ones, and privately malloc()-ed pools.
Thus introduce annotations for the memory types supported and change
pa_mempool_new() into a factory method based on required memory.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
In future commits, server-wide SHMs will be replaced with per-client
ones that will be dynamically created and freed according to clients
connections open and close.
Meanwhile, current PA design does not guarantee that the per-client
mempool blocks are referenced only by client-specific objects.
Thus reference count the pools and let each memblock inside the pool
itself, or just attached to it, increment the pool's refcount upon
allocation. This way, per-client mempools will only be freed when no
further component in the system holds any references to its blocks.
DiscussionLink: https://goo.gl/qesVMV
Suggested-by: Tanu Kaskinen <tanuk@iki.fi>
Suggested-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
This will likely be needed in the future when we start supporting high
bitrate passthrough, and there actually seem to be people 352/384 kHz
out there (potentially as an intermediate production step).
We might be compiled without eventfd support, or something else
might go wrong. And it's fully possible to continue using the old
channel rather than just disconnecting.
Signed-off-by: Pierre Ossman <ossman@cendio.se>
I want to enable client.conf.d, because in OpenEmbedded-core we have
a graphical environment called Sato that runs as root. Sato needs to
set allow-autospawn-for-root=true in client.conf, but the default
configuration in OpenEmbedded-core should not set that option. With
this patch, I can create a Sato-specific package that simply installs
50-sato.conf in /etc/pulse/client.conf.d without conflicting with the
main client.conf coming from a different package.
daemon.conf.d is enabled just because it would be strange to not
support it while client.conf.d is supported.
This allows a configuration scheme where after loading configuration
from "somefile", the parser loads configuration from files in
directory "somefile.d". This feature needs to be enabled on a per-file
basis, though, and this patch doesn't yet enable the feature for any
files.
From the NetBSD manual:
The first argument of these functions is of type int, but only a very
restricted subset of values are actually valid. The argument must either
be the value of the macro EOF (which has a negative value), or must be a
non-negative value within the range representable as unsigned char.
Passing invalid values leads to undefined behavior.
-- ctype(3)
The gnome/unity-control-center UIs have a master volume slider, and
three sub-sliders: balance, fade, and subwoofer. Balance and fade
use PA's set_balance and set_fade APIs accordingly, but the subwoofer
slider sometimes does unintuitive things.
In order to make that slider behave better, let's add a LFE balance
API that these volume control UIs can use instead. With this API,
the UI can balance between "no subwoofer" and "only subwoofer" with
"equal balance" in the middle, which would make it more consistent
with the behaviour of the other sliders.
BugLink: https://bugzilla.gnome.org/show_bug.cgi?id=753847
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Seccomp-BPF uses SIGSYS signal to trigger
the trap handler attached to sys_open.
If the signal is blocked then the kernel kills
the process whenever pulse audio calls 'open'.
The result backtrace is terminating in sys_open.
That's why it is required to keep SIGSYS unblocked
if it is currently unblocked and trapped.
This patch allows to have pulse audio working
in the Chromium sandbox.
Signed-off-by: Julien Isorce <j.isorce@samsung.com>
Signed-off-by: Arun Raghavan <git@arunraghavan.net>
Commit 262bdae0330e used symbols which are only available if systemd
support was compiled in. Fix by using the appropriate #ifdef guards.
Also document the resulting PULSE_LOG_JOURNAL environment variable
behavior if systemd journal support was not compiled in.
[Diwic: changed wording slightly.]
Reported-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
By introducing such an environment variable, applications using the
PA client libraries can configure these libraries to send their logs
directly to the journal.
While client libraries journal logging can be indirectly achieved
using PULSE_LOG_SYSLOG, this pollutes the journal. Meta data gets
replicated twice: once in the journal meta fields and once in the
syslog(3) plain-text message itself.
For attaching any backtraces, also introduce the PA-specific journal
meta field PULSE_BACKTRACE. This is the recommend journal practice
instead of appending any furuther data to the logging message itself.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Document how to modify the client libraries logging behvaior
using any of the PA-specific environment variables.
Using the PULSE_LOG_* environment variables makes debugging
and tracing PA applications quite easy, thus the need for an
official documentation text.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
In case the sample spec is not known, as can be the case when
pa_stream_new_extended is used, we cannot satisfy the PULSE_LATENCY_MSEC
request.
As a workaround disable being able to use PULSE_LATENCY_MSEC in this case.
Reported-by: Fritsch <fritsch@xbmc.org>
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
pa_tagstruct_new() is called either with no data, i.e. (NULL, 0)
to create a dynamic tagstruct or with a pointer to fixed data
introduce a new function pa_tagstruct_new_fixed() for the latter case
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
see
https://bugs.freedesktop.org/show_bug.cgi?id=85011
in case NOAUTOSPAWN is set and no server has been specified, PA starts listening on DBUS
for a new server, and the state is PA_CONTEXT_CONNECTING, but pa_context_connect()
returns -1; it should return 0.
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
FSF addresses used in PA sources are no longer valid and rpmlint
generates numerous warnings during packaging because of this.
This patch changes all FSF addresses to FSF web page according to
the GPL how-to: https://www.gnu.org/licenses/gpl-howto.en.html
Done automatically by sed-ing through sources.
This reverts commit 7276faca72.
Using the new systemd socket activation for PulseAudio will cause PulseAudio to not
have any connection with D-Bus, breaking device reservation protocol, module-jackdbus-detect
and module-dbus-protocol. Therefore, autospawn is now still enabled by default even if you
build with systemd daemon headers.
When enabled, this method is prefered over pulseaudio's built in
systems so we should try our best to ensure that it cannot be spawned
outside of the mechanisms desired.
Packagers should call 'systemctl --global enable pulseaudio.socket' to
enable the socket for all users, or alternatively ship an enabling
symlink in /usr/lib/systemd/user/sockets.target.wants/ folder. It may
also make sense for distributions to add in a ConditionNNN= line to the
socket unit if they have a downstream mechanism for enabling or
disabling pulseaudio.
If individual users wish to opt out of this vendor (or administrator)
decision, they can call 'systemctl --user mask pulseaudio.socket'
If somebody tries to push a non-frame-aligned memblock onto the
memblockq, then we should fail the write. Otherwise the daemon will
crash, see https://bugs.freedesktop.org/show_bug.cgi?id=77595
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
This fixes getting the binary name in the Hurd, or any other port using
the GNU C library, but only in the case where the library is directly
linked to. Opening with dlopen will not work.
Change in v3: reorder header includes and definitions
Change in v2: use a weak reference to main, so that we
don't crash when main cannot be found.
New function allows to pass data pointer that is a member
of the outer structure that need to be freed too when data
is not needed anymore.
Signed-off-by: Lukasz Marek <lukasz.m.luki2@gmail.com>
Since we don't have "limited" clients, a client that authenticates
correctly is automatically authorized. However, it's the authentication
that can go wrong, rather than the authorization.
Buglink: https://bugs.freedesktop.org/show_bug.cgi?id=78566
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Debug and info messages are primarily meant for developers,
rather than end users. Let's save translators' time,
and leave them untranslated.
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
json_tokener_parse() simply returns NULL on error these days
latest json-c (post 0.12) doesn't automatically include json-c/bits.h anymore
causing compilation errors
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
pulse/format.c: In function 'pa_format_info_get_prop_type':
pulse/format.c:252:5: warning: implicit declaration of function 'is_error' [-Wimplicit-function-declaration]
pulse/format.c:287:13: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
pulse/format.c:293:13: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
pulse/format.c: In function 'pa_format_info_get_prop_int_range':
pulse/format.c:364:5: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
pulse/format.c:369:5: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
pulse/format.c: In function 'pa_format_info_prop_compatible':
pulse/format.c:676:9: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
pulse/format.c:680:9: warning: 'json_object_object_get' is deprecated (declared at /usr/local/include/json-c/json_object.h:290) [-Wdeprecated-declarations]
json-c 0.10 (released 20120530) added json_object_object_get_ex()
json-c 0.12 (released 20140410) deprecated json_object_object_get()
PulseAudio depends on json 0.9 or json-c 0.11, drop support for json 0.9
in a subsequent patch and require json-c 0.11 (this will also avoids confusion
which json package is needed due to the upstream rename)
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
json-c documentation states that "No reference counts will be changed.
There is no need to manually adjust reference counts through the
json_object_put/json_object_get methods unless..."
hence fix pa_format_info_get_prop_type() and pa_format_info_get_prop_int_range();
note that pa_format_info_prop_compatible() is OK
the json_object_array_get_idx() bug reported by Arun, thanks!
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Cc: Arun Raghavan <arun@accosted.net>
handle_srbchannel_memblock() should return when memblock sanity checks fail
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Cc: David Henningsson <david.henningsson@canonical.com>
This comment can potentially save a lot of debugging effort and fixing
an ABI break, even though I don't think it's particularly likely that
anyone will ever extend pa_ext_device_manager_role_priority_info.
A recent patch broke the build on FreeBSD, which does not have
HAVE_CREDS defined. Also, make sure any attempts to enable the
srbchannel on such architectures fail.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=80642
Reported-by: Ryan Lortie
Signed-off-by: David Henningsson <david.henningsson@canonical.com>
