This commit cause bug - no sound after end voice call when
using speaker. Sound come back after audioflinger send
standby and hal reenable speaker for new stream (usecase).
Reason - hal need reenable speaker (or other devices) for
other streams (usecases).
This reverts commit f33a728b205f7656ca6211a07e3dd69c623b75ca.
Change-Id: Icde4e83ab824008e931b681cdbb6e11d6a2a7aa5
Use AUDIO_DEVICE_OUT_BLUETOOTH_SCO
AUDIO_DEVICE_OUT_BLUETOOTH_SCO_HEADSET and
AUDIO_DEVICE_OUT_BLUETOOTH_SCO_CARKIT
and not only AUDIO_DEVICE_OUT_BLUETOOTH_SCO_HEADSET for voice calls.
Remove default voice_mic for sco devices - why we need default mic
for sco output devices if now we handle all sco output devices.
Change-Id: Ibf4eb42eebcf7dc229649d8afe6e8b3ea9354165
Myabe this related to htc flounder hw but this
not needed for samsung and this cause bug when
hal trying open bt pcm device for system sounds
while voice call active on sco and bt pcm device
opened with start_voice_session_bt_sco.
Change-Id: I75482dad42feccae5e8ca9df2623ef4693a7bdb0
Add check_and_route_usecases -
this function is to make sure that all the usecases that are active on
the hardware codec backend are always routed to any one device that is
handled by the hardware codec.
Example:
When voice call starting - we need earpice -
hal enable earpice, but befor - we put call button -
and hal enable speaker and play sound - "click"
And now we have speaker and earpice enabled.
Later when audioflinger send standby (default after 3 sec) to stream
which use speaker - hal disable speaker while voice call active - and
we have no sound in earpice.
check_and_route_usecases will not allow such situations.
Change-Id: Ibc534bb14c25e4a4ffb4f3487d1424e9e1ed6a67
This never use - because -
out->sample_rate and out->config configured in adev_open_output_stream
out->config = can have deep or no pcm config with only 48000 sample rate.
Change-Id: Ie1c4025585ab6e1b7a900486181c1bceeff88139
This is part of a cleanup series.
In the future, we should reconsider the struct usage and (maybe?) move
the members into one common place.
Change-Id: I3218d75acb0c4965332ba0de30b40ca7821596a7
Some devices like the Samsung Galaxy S6, require this
to be run at the end of the voice session. However some
devices also need this to be run at the start of the
voice session whereas the S6 does not (see previous commit).
Change-Id: I833b57e262795fcbd8be6f44afb17b0f1ca6fee3
Signed-off-by: Usaamah Patel <ussyp2015@gmail.com>
Fixes: BUGBASH-302
Correctly use 16khz for the WBS-SCO pcm_config.
No more garbled or mute in-call sound on bt.
Apply together with a device specific port of:
https://review.lineageos.org/#/c/177996/
and REMOVE any definitions of
SCO_WBS_SAMPLE_RATE
https://review.lineageos.org/#/c/178371/
This requires testing for all devices using:
android_hardware_samsung
(universal audio HAL part)
Change-Id: I84fb1b234bf2880bf98bf99f1e4ff6c05500b0e3
Use strlcpy instead of strncpy when copying strings to make sure
the copy is always null-terminated.
Change-Id: I12d4883c22a180e2136dc8c85bc0394ddcdcb706
guarantee non-null function pointer before rild register its hidl service
Test: Build and local test
Bug: 78886552
Change-Id: I1137cc0f4f09f0304662a2227b84c2b5b228d45d
OEM hook is deprecated, so adding a way to disable it to
this radio implementation.
Bug: 75322118
Test: boot device w/ DISABLE_RILD_OEM_HOOK works, lshal
Change-Id: Ie7ade48476d2c330df608e9cc8dab805f84dd81d
received.
If cached value for NITZ is used, the time at which it was
received needs to be cached too.
Test: Basic telephony sanity
Bug: 72283604
Change-Id: I8f443171c4583e3eab9be7973d7714ae6c7ab6af
service and add null-checks.
The payload of a SAP request could be freed twice in certain scenarios.
Also, add null-checks to prevent dereferencing of null pointers.
Bug: 64729356
Test: Manually run the fuzz tests and ensure that there is no crash in
rild
Change-Id: Ib7ae269fa5297d6acea267337b220b8858c82bae
RIL_RadioFunctions
Re-add constant for function pointer table to
RIL_RadioFunctions. Const was accidentally removed
during a warnings cleanup.
Bug: 64231560
Test: compilation
Change-Id: I93675980c51d9162f624495d12fe3c97f1da8031
clang is the default compiler since Android nougat
Change-Id: I0555fa8067352fd43b47aabe80be1224597da17e
Signed-off-by: Lennart Wieboldt <lennart.1997@gmx.de>
Some new S7 devices are sold with cc:c0:79 MAC Address.
Adding this address in file macaddr_mappings.h is required in order to load proper drivers.
Without it, the WiFi does not work properly : signal is weak
Change-Id: I9845f8739cf3e5a4caea9eeb989576877d088731
(cherry picked from commit f66c95bcae5f4a0453b5c5f035f02e476145667b)
This allows us to distinguish not only by platform and subtype at build
time, but also by device-specific model.
It is required because two devices may share the same kernel sources and
thus build the same dtbs, but the resulting dt.img should be devie
specific and only contain the specific dtbs for that device.
Otherwise we are likely to violate dt.img size constraints by including
too many dtbs (which is the case for a5y17lte and a7y17lte right now).
For now, the feature will be optional (existing samsung_dtbh.h headers
don't have to be changed) but that might change in the future.
Change-Id: I2d804f9f1a14051e60f4de4f3d2d4f11c9b3c23f
(cherry picked from commit 2db58d0208abf3dc0f1a7c23ff34a60bcda91db1)
This change sets LOCAL_SDK_VERSION for all packages where
this is possible without breaking the build, and
LOCAL_PRIVATE_PLATFORM_APIS := true otherwise.
Setting one of these two will be made required soon, and this
is a change in preparation for that. Not setting LOCAL_SDK_VERSION
makes the app implicitly depend on the bootclasspath, which is
often not required. This change effectively makes depending on
private apis opt-in rather than opt-out.
Test: make relevant packages
Bug: 73535841
Change-Id: Id4757bcca09eda716e0a1098d9290928fa87f22d
New herolte OUI observed: 1c:70:22
Checked on hwaddress.com
Change-Id: I4a2c15e315b46af81f815114cf85f0d1cced87f3
(cherry picked from commit ba6c738c2f39fc1011ae5f61067f84b1bf430111)
* Sometimes, the modem is sending 1-2 extra fields with
the country mcc, which confuses ServiceStateTracker
* Drop the extra data here, instead of in our RIL class
[haggertk]: Forward port to ril-caf on lineage-15.
Change-Id: Ifbec67bb0dac271226bd8b5471deaf6a2ef33f2b
Current get_pcm_device return always first finded pcm_profile.
And out->pcm_dev_list always have only pcm_device_playback pcm_profiles.
But if we need use deep_buffer pcm device
we need use deep_bufer pcm_config - current hal
don't did this...
And we configure out->config in adev_open_output_stream
function with proper pcm_config (deep or no).
Change-Id: I86fdb18202afb0e83393f32830d57c929f174ae6
(cherry picked from commit 39d1d019965d04ddfad2c643de53cd5501803aa4)
stenkinevgeniy