Age | Commit message (Collapse) | Author | Files | Lines |
|
As the "controller" isn't any more in a submodules
but in a library, it is required to add the dependency
int he recipe.
Bug-AGL: SPEC-2139
Change-Id: Ia629faf627310ee8ae9de363fd1fbc30f4c55f31
Signed-off-by: José Bollo <jose.bollo@iot.bzh>
|
|
The verb 'power' is now replaced by 'adapter_state'. The response
object is also different (no compatibility with older API).
Bug-AGL: SPEC-1630
Change-Id: I2e08c33e2936d93cfdec0c498570c77177a4b81e
Signed-off-by: Stéphane Desneux <stephane.desneux@iot.bzh>
|
|
Bug-AGL: SPEC-1505
- Remove UNICENS legacy hal from build until the new hal plugin
is available
- Remove ldpath which was required to load the legacy hal
(avoids binder startup error since the path is no longer available)
Change-Id: If25094d4f944e887fed0ed10c5cae8049ce6db82
Signed-off-by: Tobias Jahnke <tobias.jahnke@microchip.com>
|
|
workaround)
This patch introduces a wait loop before starting the 4A binder:
* wait for bluetooth module to be loaded
* wait for bluetooth controller to be up & running
* wait for bluetooth manager to be initialized
Then 4A can start with BT-audio support.
If no BT controller is detected in 5secondes, the loop exits and
4A doesn't initialize BT-Audio
Bug-AGL: SPEC-1762, SPEC-1763
Change-Id: Ibe851dee871d183511c7bca03411b35ed023f422
Signed-off-by: Stephane Desneux <stephane.desneux@iot.bzh>
|
|
Since the integration of bluez-alsa support in softmixer,
and as warned by the readme of bluez-alsa, there are potential
(and effective !) deadlocks in libasound. These deadlocks come
especially when uing ioplug PCMs.
This hack disables the calls to snd_pcm_lock/unlock in libasound.
Notice that the design that consists in using such PCM plugs
for communicating with the bluealsa daemon is just a short term
solution. When the direct support of bluezalsa communication
is integrated in the softmixer, that hack should be removed.
Change-Id: Ib6376fd2e145b3240ab1a68b9c760ff9e4c063c4
Signed-off-by: Thierry Bultel <thierry.bultel@iot.bzh>
|
|
This bump is based on recipes coming for eel branch (revision e029270):
In 4A layer, we now have:
* 4a-softmixer: SRCREV = "731c21e80ba45339b7dd9b8eae63accd3597d281"
* 4a-hal-unicens: SRCREV = "${AGL_APP_REVISION}"
* agl-service-audio-4a: SRCREV = "${AGL_APP_REVISION}"
* 4a-hal-generic: SRCREV = "8142c7ed42167dff7e526952f91909d679453251"
* 4a-mixer: SRCREV = "${AGL_APP_REVISION}"
* 4a-alsa-core: SRCREV = "${AGL_APP_REVISION}"
Change-Id: Ib5708e9551f0646e0206deac8795f3b09f5a0569
Signed-off-by: Stephane Desneux <stephane.desneux@iot.bzh>
|
|
* Update Policy level of each pcm
* Fix Pulseaudio start
* Add 4A module config to pulse audio
SPEC-1140: we need to have orderer PCM list in file 90-4a-modules.pa,
to fix pulseaudio startup. (From lower to higher).
Bug-AGL: SPEC-1136
Bug-AGL: SPEC-1103
Change-Id: I847711d9076a0d9817176a5cc3dee823476d8703
Signed-off-by: ronan <ronan@iot.bzh>
|
|
Recipes need to use AGL_APP_REVISION (for apps) and AGL_DEFAULT_REVISION (for non-apps)
respectively instead of plain AUTOREV.
AGL_APP_REVISION / AGL_DEFAULT_REVISION will default to AUTOREV in normal operation until
a release is cut in which case we will put in the tag.
See conf/distro/poky-agl.conf
Change-Id: Ib346e9c9bd118d245cccbec3c4bbd3c99e1ec457
Signed-off-by: Jan-Simon Möller <jsmoeller@linuxfoundation.org>
|
|
Change-Id: Ic949f35e0f4fa29fc24ae69b443965f8f1ac99f8
Signed-off-by: José Bollo <jose.bollo@iot.bzh>
|
|
The new version of the framework makes more use of
system units. This patch has 2 effects:
- it doesn't stop the system forever in case of error
implying the lose of the awaited files
- it accepts the service file to reside in either system
or user scope
Change-Id: I7bd8a38811a9f0dccddcf922a330ed15b652ccc3
Signed-off-by: José Bollo <jose.bollo@iot.bzh>
|
|
This is a move of the work done initially by Ronan Le Martret <ronan.lemartret@iot.bzh>
in https://gerrit.automotivelinux.org/gerrit/#/c/11651/
Change-Id: I3c2124b186e72fd9804ff713b22a36a93283650e
Signed-off-by: Stephane Desneux <stephane.desneux@iot.bzh>
|