Hi Arun,
Arun Raghavan wrote:
> Hi Hajime,
> I'm still working on this, it's slow going since I have other things to
> do in parallel.
Thank you for your reply. Yeah that's totally fine. Just wanted make sure if it
was appropriate to submit a new series.
> It's a bit hard to review this work
On Fri, Feb 12, 2016 at 02:09:31AM +0200, Ahmed S. Darwish wrote:
> The PA daemon currently uses a single SHM file for all the clients
> sending and receiving commands over the low-latency srbchannel
> mechanism.
>
> To safely run PA daemon in system mode later using memfds, and to
> provide the
On Wed, 2016-02-10 at 09:56 +0530, Arun Raghavan wrote:
> On 9 February 2016 at 23:41, Tanu Kaskinen wrote:
> > On Mon, 2016-01-18 at 13:06 +0530, a...@accosted.net wrote:
> > > @@ -387,7 +392,8 @@ void pa_webrtc_ec_record(pa_echo_canceller *ec, const
> > > uint8_t *rec, uint8_t
On Mon, 2016-01-18 at 13:06 +0530, a...@accosted.net wrote:
> From: Arun Raghavan
>
> Needed for upcoming beamforming code.
> ---
> src/modules/echo-cancel/echo-cancel.h | 2 +-
> src/modules/echo-cancel/webrtc.cc | 14 --
> 2 files changed, 9
On 11.02.2016 12:34, Tanu Kaskinen wrote:
> On Wed, 2016-02-10 at 17:37 +0100, Stefan Neufeind wrote:
>> Hi,
>>
>> upon pairing a phone I automatically get an A2DP source. And when
>> playing sound from that phone it plays back on the PC through an
>> autocreated loopback.
>>
>> However the
On 22.01.2016 04:06 PM, Tanu Kaskinen wrote:
> On Mon, 2016-01-18 at 13:06 +0530, a...@accosted.net wrote:
>> From: Arun Raghavan
>>
>> The code now needs C++11 support to compile with the updated
>> webrtc-audio-processing library.
>> ---
>> configure.ac
On Wed, 2016-02-10 at 17:37 +0100, Stefan Neufeind wrote:
> Hi,
>
> upon pairing a phone I automatically get an A2DP source. And when
> playing sound from that phone it plays back on the PC through an
> autocreated loopback.
>
> However the loopbacked sound comes out of the wrong pair of
(I found this from the moderation queue.)
On Mon, 2016-02-08 at 19:12 +, J Copse wrote:
> I seem to be unable to get the captcha to load on the subscription
> page. pulseaudio-discuss Info Page
The info page now says "We're trying to fix this; spammers are abusing
the machine" where the
Fix improper unducking behavior when using this module with "global=1" argument.
BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=93994
Signed-off-by: Sangchul Lee
---
src/modules/module-role-ducking.c | 18 +-
1 file changed, 17 insertions(+), 1
Hello!
The all-improved memfd patch series ;-)
==> v1 cover letter:
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
Soon we're going to have three types of memory pools: POSIX shm_open()
pools, memfd memfd_create() ones, and privately malloc()-ed pools.
Due to such diversity of mempool types, transform pa_mempool_new()
into a factory method that returns the required type of memory pool
according to given
Mempools use "pa_shm" for their pools memory allocation. pa_shm is
then overloaded with two unrelated tasks: either allocating the pool
using posix SHM or allocating the same pool using private malloc()s.
The choice depends on system configuration and abilities.
This will not scale when a third
Now that we have the necessary infrastructure for memexporting and
mempimporting a memfd memblock, extend that support higher up in the
chain with pstreams.
A PulseAudio endpoint can now _transparently_ send a memfd memblock
to the other end by simply calling pa_pstream_send_memblock().
If the
Now that all layers in the stack support memfd blocks, use memfd
pools for client context and audio data by default.
Signed-off-by: Ahmed S. Darwish
---
man/pulse-client.conf.5.xml.in | 5 +
src/pulse/client-conf.c| 1 +
src/pulse/client-conf.h| 2 +-
The PA daemon currently uses a single SHM file for all the clients
sending and receiving commands over the low-latency srbchannel
mechanism.
To safely run PA daemon in system mode later using memfds, and to
provide the necessary ground work for sandboxing, create the
srbchannel SHM files on a
The daemon `shm-size-bytes' configuration value was read, and then
directly used, for creating the initial server-wide SHM files.
This is fine for now, but soon, such server-wide SHMs will be replaced
with per-client SHM files that will be dynamically created and deleted
according to clients open
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
Add support for transfering memfd-backed blocks without passing their
file descriptor every time, thus minimizing overhead and the
possibility of fd leaks.
To accomplish this, a new type of 'permanent' segments is introduced.
Suggested-by: David Henningsson
Color global mempools with a special mark. Almost all pools are
now created on a per client basis except the pa_core's mempool,
which is shared between all clients.
This special marking is needed for handling memfd-backed pools.
To avoid fd leaks, memfd pools are registered with the connection
Now that all layers in the stack support memfd blocks, use memfd
pools for the global core mempool by default.
Also introduce "disable-memfd" daemon argument and configuration
option to disable memfd support at run-time, if desired.
Signed-off-by: Ahmed S. Darwish
---
Use memfd blocks for the srbchannel if:
- we have memfd support compiled in, and ..
- client supports PA protocol version >= 31, and ..
- client states that it can handle memfd shared memory
Otherwise fall-back to the regular POSIX shm_open() techniques.
Signed-off-by: Ahmed S. Darwish
21 matches
Mail list logo