profile/ivi/pulseaudio.git
15 years agovolume_sse: add sse optimisations
Wim Taymans [Wed, 12 Aug 2009 18:44:12 +0000 (20:44 +0200)]
volume_sse: add sse optimisations

15 years agovolume_mmx: fix mmx code a bit
Wim Taymans [Wed, 12 Aug 2009 18:43:37 +0000 (20:43 +0200)]
volume_mmx: fix mmx code a bit

15 years agovolume: add first mmx optimized function
Wim Taymans [Wed, 12 Aug 2009 15:08:41 +0000 (17:08 +0200)]
volume: add first mmx optimized function

Add code for an mmx optimized version of s16ne volume scaling. Install the
custom function.

15 years agosamples-util: add padding to volume array
Wim Taymans [Wed, 12 Aug 2009 15:03:30 +0000 (17:03 +0200)]
samples-util: add padding to volume array

Pad the volume array with a copy of the start. We'll need this later to be able
to write optimized functions.

15 years agosample-util: move volume code to separate file
Wim Taymans [Tue, 11 Aug 2009 15:10:44 +0000 (17:10 +0200)]
sample-util: move volume code to separate file

Move the volume code into a separate file with the reference C implementations.
Add a function to retrieve the volume function and one to install a new one.

15 years agosample-util: move some functions around
Wim Taymans [Tue, 11 Aug 2009 14:43:46 +0000 (16:43 +0200)]
sample-util: move some functions around

Move some stuff around before splitting it into a separate file.

15 years agosample-utils: coding style cleanup
Wim Taymans [Tue, 11 Aug 2009 14:25:44 +0000 (16:25 +0200)]
sample-utils: coding style cleanup

Make the coding style match the rest of pulseaudio more.
Remove some liboil functions, they seem unoptimized and likely slower than our
handrolled versions here.

15 years agosample-utils: split out functions from case
Wim Taymans [Tue, 11 Aug 2009 13:15:57 +0000 (15:15 +0200)]
sample-utils: split out functions from case

Move the volume functions out of the switch case and use a table indexed by the
sample format to find the volume function.

15 years agoModification of the thread-mainloop doc to ensure that nobody frees the api as it...
Xabier Rodriguez Calvar [Tue, 30 Jun 2009 16:23:17 +0000 (18:23 +0200)]
Modification of the thread-mainloop doc to ensure that nobody frees the api as it is owned by the loop.

15 years agoModification of the mainloop doc to ensure that nobody frees the api as it is owned...
Xabier Rodriguez Calvar [Tue, 30 Jun 2009 16:22:44 +0000 (18:22 +0200)]
Modification of the mainloop doc to ensure that nobody frees the api as it is owned by the loop.

15 years agoModification of the glib-mainloop doc to ensure that nobody frees the api as it is...
Xabier Rodriguez Calvar [Tue, 30 Jun 2009 16:20:03 +0000 (18:20 +0200)]
Modification of the glib-mainloop doc to ensure that nobody frees the api as it is owned by the loop.

15 years agocore: check return value of getgrnam_r() instead of errno
Lennart Poettering [Wed, 19 Aug 2009 22:20:03 +0000 (00:20 +0200)]
core: check return value of getgrnam_r() instead of errno

According to POSIX getgrnam_r() returns the error code as return value,
and not in errno. Honour that.

Pointed out and inspired by a patch from Ted Percival.

15 years agoi18n: run make update-po
Lennart Poettering [Wed, 19 Aug 2009 02:16:35 +0000 (04:16 +0200)]
i18n: run make update-po

15 years agoi18n: add ko to LINGUAS
Lennart Poettering [Wed, 19 Aug 2009 02:15:36 +0000 (04:15 +0200)]
i18n: add ko to LINGUAS

15 years agoMerge commit 'origin/master-tx'
Lennart Poettering [Wed, 19 Aug 2009 02:14:00 +0000 (04:14 +0200)]
Merge commit 'origin/master-tx'

15 years agolirc: replace manual code by pa_cvolume_{inc|dec}()
Lennart Poettering [Wed, 19 Aug 2009 01:37:42 +0000 (03:37 +0200)]
lirc: replace manual code by pa_cvolume_{inc|dec}()

15 years agommkbd: replace manual code by pa_cvolume_{inc|dec}()
Lennart Poettering [Wed, 19 Aug 2009 01:37:29 +0000 (03:37 +0200)]
mmkbd: replace manual code by pa_cvolume_{inc|dec}()

15 years agosink: volume handling rework, new flat volume logic
Lennart Poettering [Wed, 19 Aug 2009 00:55:02 +0000 (02:55 +0200)]
sink: volume handling rework, new flat volume logic

- We now implement a logic where the sink maintains two distinct
  volumes: the 'reference' volume which is shown to the users, and the
  'real' volume, which is configured to the hardware. The latter is
  configured to the max of all streams. Volume changes on sinks are
  propagated back to the streams proportional to the reference volume
  change. Volume changes on sink inputs are forwarded to the sink by
  'pushing' the volume if necessary.

  This renames the old 'virtual_volume' to 'real_volume'. The
  'reference_volume' is now the one exposed to users.

  By this logic the sink volume visible to the user, will always be the
  "upper" boundary for everything that is played. Saved/restored stream
  volumes are measured relative to this boundary, the factor here is
  always < 1.0.

- introduce accuracy for sink volumes, similar to the accuracy we
  already have for source volumes.

- other cleanups.

15 years agomatch: document how broken the module-match logic is
Lennart Poettering [Wed, 19 Aug 2009 00:37:35 +0000 (02:37 +0200)]
match: document how broken the module-match logic is

15 years agovolume: introduce pa_cvolume_{inc|dec}()
Lennart Poettering [Wed, 19 Aug 2009 00:32:36 +0000 (02:32 +0200)]
volume: introduce pa_cvolume_{inc|dec}()

15 years agovolume: use PA_VOLUME_MAX instead of (pa_volume_t) -1
Lennart Poettering [Wed, 19 Aug 2009 00:31:11 +0000 (02:31 +0200)]
volume: use PA_VOLUME_MAX instead of (pa_volume_t) -1

15 years agoudev: allow passing of ignore_dB= parameter to alsa modules
Lennart Poettering [Wed, 19 Aug 2009 00:29:59 +0000 (02:29 +0200)]
udev: allow passing of ignore_dB= parameter to alsa modules

15 years agosource: rework volume handling
Lennart Poettering [Tue, 18 Aug 2009 23:35:43 +0000 (01:35 +0200)]
source: rework volume handling

- drop the 'virtual_' prefix from s->virtual_volume since we don't
  distuingish between reference and real volumes for sources

- introduce an accuracy for source volumes: if the hardware can control
  the volume "close enough" don't necessarily adjust the rest in
  software unless it is beyond a certain threshold. This should save a
  little bit of CPU at the expensive of a bit of accuracy in volume
  handling.

- other minor cleanups

15 years agomacro: extend comments a bit
Lennart Poettering [Tue, 18 Aug 2009 23:03:09 +0000 (01:03 +0200)]
macro: extend comments a bit

15 years agovoltest: extend test to verify correctness of _multiply() and _divide()
Lennart Poettering [Tue, 18 Aug 2009 23:02:32 +0000 (01:02 +0200)]
voltest: extend test to verify correctness of _multiply() and _divide()

15 years agodbus: never return DBUS_HANDLER_RESULT_HANDLED in filter callbacks, since other callb...
Lennart Poettering [Tue, 18 Aug 2009 22:59:23 +0000 (00:59 +0200)]
dbus: never return DBUS_HANDLER_RESULT_HANDLED in filter callbacks, since other callbacks might stell want to have the messages

15 years agovolume: round properly when showing human readable volume percentages
Lennart Poettering [Tue, 18 Aug 2009 22:58:20 +0000 (00:58 +0200)]
volume: round properly when showing human readable volume percentages

15 years agovolume: simplify volume multiplifactions, do them in integer only
Lennart Poettering [Tue, 18 Aug 2009 22:57:58 +0000 (00:57 +0200)]
volume: simplify volume multiplifactions, do them in integer only

15 years agovolume: introduce pa_cvolume_min() and pa_cvolume_min_mask()
Lennart Poettering [Tue, 18 Aug 2009 22:56:16 +0000 (00:56 +0200)]
volume: introduce pa_cvolume_min() and pa_cvolume_min_mask()

15 years agomacro: simplify page/word alignment macros a bit
Lennart Poettering [Mon, 17 Aug 2009 17:17:10 +0000 (19:17 +0200)]
macro: simplify page/word alignment macros a bit

15 years agomacro: add PA_ROUND_UP/PA_ROUND_DOWN macros
Lennart Poettering [Mon, 17 Aug 2009 17:15:33 +0000 (19:15 +0200)]
macro: add PA_ROUND_UP/PA_ROUND_DOWN macros

15 years agobluetooth: fix match syntax
Lennart Poettering [Mon, 17 Aug 2009 17:14:35 +0000 (19:14 +0200)]
bluetooth: fix match syntax

15 years agobluetooth: make NameOwnerChanged filter matches more focussed
Lennart Poettering [Mon, 17 Aug 2009 01:51:17 +0000 (03:51 +0200)]
bluetooth: make NameOwnerChanged filter matches more focussed

15 years agovolume: document when arguments of certain functions may overlap
Lennart Poettering [Mon, 17 Aug 2009 01:50:43 +0000 (03:50 +0200)]
volume: document when arguments of certain functions may overlap

15 years agolibpulse: some minor optimizations when checking equality
Lennart Poettering [Mon, 17 Aug 2009 01:50:04 +0000 (03:50 +0200)]
libpulse: some minor optimizations when checking equality

15 years agochannelmap: minor doxygen fix
Lennart Poettering [Mon, 17 Aug 2009 01:48:51 +0000 (03:48 +0200)]
channelmap: minor doxygen fix

15 years agochannelmap: document where the WAVEX channelmap is documented
Lennart Poettering [Mon, 17 Aug 2009 01:43:05 +0000 (03:43 +0200)]
channelmap: document where the WAVEX channelmap is documented

15 years agochannelmap: adjust RFC3551 channel maps to follow spec more closely
Lennart Poettering [Mon, 17 Aug 2009 01:42:34 +0000 (03:42 +0200)]
channelmap: adjust RFC3551 channel maps to follow spec more closely

15 years agoreserver: update from upstream git
Lennart Poettering [Mon, 17 Aug 2009 01:41:14 +0000 (03:41 +0200)]
reserver: update from upstream git

15 years agovolume: add pa_cvolume_merge() call
Lennart Poettering [Mon, 17 Aug 2009 01:40:36 +0000 (03:40 +0200)]
volume: add pa_cvolume_merge() call

15 years agoSending translation for Czech
pmkovar [Mon, 17 Aug 2009 00:19:25 +0000 (00:19 +0000)]
Sending translation for Czech

15 years agoaupdate: implicitly call _write_swap() if it wasn't called explicitly
Lennart Poettering [Sun, 16 Aug 2009 00:02:51 +0000 (02:02 +0200)]
aupdate: implicitly call _write_swap() if it wasn't called explicitly

15 years agocore: add to FIXMEs
Lennart Poettering [Sat, 15 Aug 2009 22:45:56 +0000 (00:45 +0200)]
core: add to FIXMEs

15 years agomodules: add various checks to avoid selecting objects that are not linked or in...
Lennart Poettering [Sat, 15 Aug 2009 22:45:23 +0000 (00:45 +0200)]
modules: add various checks to avoid selecting objects that are not linked or in another unclear state

15 years agobluetooth: move installation of mq's earlier to avoid context asserts to be triggered
Lennart Poettering [Sat, 15 Aug 2009 22:43:22 +0000 (00:43 +0200)]
bluetooth: move installation of mq's earlier to avoid context asserts to be triggered

15 years agobluetooth: ask first for Headset and AudioSink properties, followed by Audio
Lennart Poettering [Sat, 15 Aug 2009 20:25:53 +0000 (22:25 +0200)]
bluetooth: ask first for Headset and AudioSink properties, followed by Audio

15 years agobluetooth: recognize only those BT devices that implement both the Audio and either...
Lennart Poettering [Sat, 15 Aug 2009 20:25:21 +0000 (22:25 +0200)]
bluetooth: recognize only those BT devices that implement both the Audio and either AudioSink or Headset interfaces

15 years agobluetooth: drop PA_BT_AUDIO_STATE_LAST since it is unused and we normally call that...
Lennart Poettering [Sat, 15 Aug 2009 20:24:13 +0000 (22:24 +0200)]
bluetooth: drop PA_BT_AUDIO_STATE_LAST since it is unused and we normally call that _MAX anyway

15 years agothread-mq: do final q flush only when we aren't dispatching anyway
Lennart Poettering [Sat, 15 Aug 2009 20:23:42 +0000 (22:23 +0200)]
thread-mq: do final q flush only when we aren't dispatching anyway

15 years agoasyncmsgq: introduce pa_asyncmsgq_dispatching()
Lennart Poettering [Sat, 15 Aug 2009 20:22:57 +0000 (22:22 +0200)]
asyncmsgq: introduce pa_asyncmsgq_dispatching()

15 years agoladspa/remap: sync latency flags from master sink when moving between sinks
Lennart Poettering [Sat, 15 Aug 2009 01:42:16 +0000 (03:42 +0200)]
ladspa/remap: sync latency flags from master sink when moving between sinks

15 years agocore: introduce pa_{sink|source}_update_flags()
Lennart Poettering [Sat, 15 Aug 2009 01:41:13 +0000 (03:41 +0200)]
core: introduce pa_{sink|source}_update_flags()

15 years agopulse: even in case of record stream, let's initialize req_bytes to 0
Marc-André Lureau [Mon, 20 Jul 2009 10:53:17 +0000 (13:53 +0300)]
pulse: even in case of record stream, let's initialize req_bytes to 0

15 years agobluetooth: fix typo with service_{read,write}_type
Marc-André Lureau [Mon, 20 Jul 2009 10:44:27 +0000 (13:44 +0300)]
bluetooth: fix typo with service_{read,write}_type

15 years agobluetooth: don't call pa_sink_render with 0 bytes request
Marc-André Lureau [Mon, 6 Jul 2009 14:50:51 +0000 (17:50 +0300)]
bluetooth: don't call pa_sink_render with 0 bytes request

15 years agodatabase: simple hashmap based database implementation
Juho Hämäläinen [Wed, 12 Aug 2009 15:30:14 +0000 (18:30 +0300)]
database: simple hashmap based database implementation

15 years agocombine: determine sample parameters of combined sink from underlying sinks
Maarten Bosmans [Fri, 14 Aug 2009 23:35:53 +0000 (01:35 +0200)]
combine: determine sample parameters of combined sink from underlying sinks

http://pulseaudio.org/ticket/521

15 years agocombine: warn when the latency of a stream gets too high
Lennart Poettering [Fri, 14 Aug 2009 23:16:57 +0000 (01:16 +0200)]
combine: warn when the latency of a stream gets too high

15 years agocombine: big rework
Lennart Poettering [Fri, 14 Aug 2009 23:07:37 +0000 (01:07 +0200)]
combine: big rework

15 years agocombine: drop adjust_timestamp variable because it is unused
Lennart Poettering [Fri, 14 Aug 2009 23:04:21 +0000 (01:04 +0200)]
combine: drop adjust_timestamp variable because it is unused

15 years agoladspa: name sink after human readable plugin name, not the id string
Lennart Poettering [Fri, 14 Aug 2009 23:02:16 +0000 (01:02 +0200)]
ladspa: name sink after human readable plugin name, not the id string

15 years agomodule-ladspa: allow moving of sink, forward fixed latency
Lennart Poettering [Fri, 14 Aug 2009 23:01:52 +0000 (01:01 +0200)]
module-ladspa: allow moving of sink, forward fixed latency

15 years agomodule-remap: allow moving of sink, forward fixed latency
Lennart Poettering [Fri, 14 Aug 2009 22:59:26 +0000 (00:59 +0200)]
module-remap: allow moving of sink, forward fixed latency

This is a bigger change reworking a number of things:

- We now allow moving of the remap sink betwween backend sinks like any
  other stream.

- We forward the fixed latency parameter of the underlying sinks the
  same way as the dynamic latency.

15 years agoladspa: move LADSPA_Data size check to compile time
Lennart Poettering [Fri, 14 Aug 2009 22:58:19 +0000 (00:58 +0200)]
ladspa: move LADSPA_Data size check to compile time

15 years agoremap: unify argument order with other modules
Lennart Poettering [Fri, 14 Aug 2009 22:57:36 +0000 (00:57 +0200)]
remap: unify argument order with other modules

15 years agohook-list: make use of PA_LLIST_FOREACH
Lennart Poettering [Fri, 14 Aug 2009 22:55:31 +0000 (00:55 +0200)]
hook-list: make use of PA_LLIST_FOREACH

15 years agonative-protocol: downgrade volume change log messages
Lennart Poettering [Fri, 14 Aug 2009 22:54:51 +0000 (00:54 +0200)]
native-protocol: downgrade volume change log messages

15 years agocore: document difference between IO and main thread view on requested latency
Lennart Poettering [Fri, 14 Aug 2009 22:54:25 +0000 (00:54 +0200)]
core: document difference between IO and main thread view on requested latency

15 years agocore: don't update latency range if not changed
Lennart Poettering [Fri, 14 Aug 2009 22:54:02 +0000 (00:54 +0200)]
core: don't update latency range if not changed

15 years agocore: call pa_sink_get_latency_within_thread() instead of going directly via process_...
Lennart Poettering [Fri, 14 Aug 2009 22:52:50 +0000 (00:52 +0200)]
core: call pa_sink_get_latency_within_thread() instead of going directly via process_msg()

15 years agocore: make fixed latency dynamically changeable
Lennart Poettering [Fri, 14 Aug 2009 22:48:14 +0000 (00:48 +0200)]
core: make fixed latency dynamically changeable

This of course makes the name 'fixed' a bit of a misnomer. However the
definitions are now like this:

fixed latency: the latency may change during runtime, but is solely
controlled by the backend, the client has no influence.

dynamic latency: the latency may change during runtime, influenced by
the requests of the clients.

i.e. fixed vs. dynamic is from the perspective of the client.

15 years agocore: move rtpoll to thread_info sub structure
Lennart Poettering [Fri, 14 Aug 2009 22:26:00 +0000 (00:26 +0200)]
core: move rtpoll to thread_info sub structure

15 years agolog: place more rate limit invocations
Lennart Poettering [Fri, 14 Aug 2009 22:16:25 +0000 (00:16 +0200)]
log: place more rate limit invocations

15 years agorescure-stream: handle failed moves as well as dying sinks/sources
Lennart Poettering [Fri, 14 Aug 2009 22:15:18 +0000 (00:15 +0200)]
rescure-stream: handle failed moves as well as dying sinks/sources

15 years agocore: split of FAIL_ON_SUSPEND into KILL_ON_SUSPEND and NO_CREATE_ON_SUSPEND
Lennart Poettering [Fri, 14 Aug 2009 22:12:53 +0000 (00:12 +0200)]
core: split of FAIL_ON_SUSPEND into KILL_ON_SUSPEND and NO_CREATE_ON_SUSPEND

15 years agonative: handle moving() callback with NULL destination properly
Lennart Poettering [Fri, 14 Aug 2009 22:05:17 +0000 (00:05 +0200)]
native: handle moving() callback with NULL destination properly

15 years agocore: introduce pa_{sink_input|source_output}_fail_move()
Lennart Poettering [Fri, 14 Aug 2009 22:03:50 +0000 (00:03 +0200)]
core: introduce pa_{sink_input|source_output}_fail_move()

15 years agomodule-stream-restore: don't fiddle with sinks/sources/streams that are not fully...
Lennart Poettering [Fri, 14 Aug 2009 21:55:32 +0000 (23:55 +0200)]
module-stream-restore: don't fiddle with sinks/sources/streams that are not fully set up yet

15 years agovolume: guarantee dB/linear conversion is reversible
Lennart Poettering [Fri, 14 Aug 2009 18:03:30 +0000 (20:03 +0200)]
volume: guarantee dB/linear conversion is reversible

15 years agoprotocol-native: log explicitly each time a client triggers a volume change
Lennart Poettering [Fri, 14 Aug 2009 17:31:05 +0000 (19:31 +0200)]
protocol-native: log explicitly each time a client triggers a volume change

15 years agodaemon: reset gids too, not just uids
Lennart Poettering [Fri, 14 Aug 2009 17:28:18 +0000 (19:28 +0200)]
daemon: reset gids too, not just uids

15 years agoresampler: round up when estimating input/output sizes
Lennart Poettering [Fri, 14 Aug 2009 02:21:09 +0000 (04:21 +0200)]
resampler: round up when estimating input/output sizes

15 years agothread-mq: never drop queued messages for the main loop
Lennart Poettering [Fri, 14 Aug 2009 02:18:23 +0000 (04:18 +0200)]
thread-mq: never drop queued messages for the main loop

Previously we might have dropped messages from IO trheads to the main
thread. This tuend out to be problematic since this cause SHM release
messages to be lost. More visibly however this could cause playback
freezing when moving streams between sinks and removing the old sink
right away.

15 years agoasyncmsgq: add pa_asyncmsgq_flush() call
Lennart Poettering [Fri, 14 Aug 2009 02:17:48 +0000 (04:17 +0200)]
asyncmsgq: add pa_asyncmsgq_flush() call

15 years agomemblockq: add pa_memblockq_get_maxrewind() API
Lennart Poettering [Fri, 14 Aug 2009 02:16:28 +0000 (04:16 +0200)]
memblockq: add pa_memblockq_get_maxrewind() API

15 years agoshm: bump shm size limit to 1GB
Lennart Poettering [Fri, 14 Aug 2009 02:14:35 +0000 (04:14 +0200)]
shm: bump shm size limit to 1GB

15 years agopacmd: handle multi word commands in argv[] properly
Lennart Poettering [Fri, 14 Aug 2009 02:12:36 +0000 (04:12 +0200)]
pacmd: handle multi word commands in argv[] properly

15 years agoSending translation for po/ko.po
eukim [Thu, 13 Aug 2009 12:20:50 +0000 (12:20 +0000)]
Sending translation for po/ko.po

15 years agocore: add functions to query max_rewind/max_request values from streams
Lennart Poettering [Thu, 13 Aug 2009 00:17:24 +0000 (02:17 +0200)]
core: add functions to query max_rewind/max_request values from streams

15 years agocore: add assert macros for verifying calling context
Lennart Poettering [Thu, 13 Aug 2009 00:14:19 +0000 (02:14 +0200)]
core: add assert macros for verifying calling context

This adds pa_assert_io_context() and pa_assert_ctl_context() in addition
to a few related macros. When called they will fail when the current execution
context is not IO resp. not control context. (aka 'thread' context vs.
'main' context)

15 years agostart-child: clean up child environment a bit better
Lennart Poettering [Wed, 12 Aug 2009 19:43:06 +0000 (21:43 +0200)]
start-child: clean up child environment a bit better

15 years agopacmd: port pacmd from select() to poll() so that we notice writer side hangups
Lennart Poettering [Wed, 12 Aug 2009 19:42:02 +0000 (21:42 +0200)]
pacmd: port pacmd from select() to poll() so that we notice writer side hangups

15 years agomemblock: rate limit 'Pool full' message
Lennart Poettering [Wed, 12 Aug 2009 19:40:38 +0000 (21:40 +0200)]
memblock: rate limit 'Pool full' message

15 years agocontext: document why we only do minimal cleanups before the autospawn exec()
Lennart Poettering [Wed, 12 Aug 2009 19:40:12 +0000 (21:40 +0200)]
context: document why we only do minimal cleanups before the autospawn exec()

15 years agodaemon: reset scheduling priority on startup, too
Lennart Poettering [Wed, 12 Aug 2009 19:37:40 +0000 (21:37 +0200)]
daemon: reset scheduling priority on startup, too

15 years agocore-util: move personality resetting into core-util
Lennart Poettering [Wed, 12 Aug 2009 19:36:52 +0000 (21:36 +0200)]
core-util: move personality resetting into core-util

15 years agodaemon: unconditionally clean up priviliges
Lennart Poettering [Wed, 12 Aug 2009 19:35:48 +0000 (21:35 +0200)]
daemon: unconditionally clean up priviliges

15 years agoconf: invert all negative boolean configuration option
Lennart Poettering [Wed, 12 Aug 2009 18:15:28 +0000 (20:15 +0200)]
conf: invert all negative boolean configuration option

15 years agoconf: add pa_config_parse_not_bool() for parsing inverse boolean configuration options
Lennart Poettering [Wed, 12 Aug 2009 18:14:31 +0000 (20:14 +0200)]
conf: add pa_config_parse_not_bool() for parsing inverse boolean configuration options

15 years agodaemon: install D-Bus system policy file for PA system instances
Lennart Poettering [Tue, 11 Aug 2009 23:50:24 +0000 (01:50 +0200)]
daemon: install D-Bus system policy file for PA system instances

Original patch from 'mgrela'.

http://pulseaudio.org/ticket/582