* This fixes the following logspam:
E/ QTI PowerHAL: Failed to acquire lock for hint_id: 1040
E/ QTI PowerHAL: Failed to acquire lock for hint_id: 1041
E/ QTI PowerHAL: Failed to acquire lock for hint_id: 10A5
With this change, the schedutil governor
has to wait 500 microseconds to increase
frequencies and 20 milliseconds to decrease
frequency. It improves performance at a
slight cost to power.
Change-Id: Ied25fc112cbb49d70ea496e0beded612478ef897
Signed-off-by: aswin7469 <aswinas@pixysos.com>
Background tasks are not UX-critical.
Free up cores to work on more important
UX tasks.
Change-Id: I01b17f15256a95513b59fd1c7d9d6d71979cebcc
Signed-off-by: aswin7469 <aswinas@pixysos.com>
After unbalancing, move these IRQs to separate cores
to improve latency.
Change-Id: Ib1c22f9c8d8c00cf74351099e88fbb5ae70e9114
Signed-off-by: aswin7469 <aswinas@pixysos.com>
Reduce the per-process limit of the mlock() family of system calls to
64 KB on account of a new CTS requirement.
Change-Id: I211668956bb6098e87cb7ef68bb7f1510d93d76f
The per-process limit for the mlock() family of system calls only
needs to be set in one group of config files. Remove from here as it
is being set in init.qcom.rc .
Change-Id: Ia7d46800eabe2954e3e2618a38ff180c88fee348
needed for using EdgeTpu.
Due to this code goes to AOSP, please see CL details and test result in
b/289097511#comment24.
Bug: 289097511
Change-Id: Ie0969309346cd85b4bb8ac71860a529710c73345
While implementing the capo nanoapp for audio configuration, we met
some chre socket connected error due to the access group denied.
Add context_hub group for audio hal to connect chre socket.
go/capo-algo
Bug: 141128522
Bug: 149069556
Test: FULL build with audio hal part and local prebuilts nanoapp.
Test: Audio HAL communicates normally with CHRE socket.
Change-Id: Iea84411682f4c3e08f8b37a5b21818b0e9b04983
The health AIDL HAL service provides functionalities of charger,
therefore system charger at /system/bin/charger is deprecated.
On top of that, QTI health AIDL HAL service enables suspend by
default, the equivalent of setting ro.charger.enable_suspend
for legacy charger.
Change-Id: I59c23e7974cea1174b0161f31a535fa3afa1e5c9
No longer needed now init listens for property changes on a
separate thread.
Bug: 186580823
Test: Cuttlefish boots successfully
Change-Id: Ief3a546d2154231d18957c2f703bd128966d7288
This is necessary for correct initialization of the network when the device starts. Without this, starting the mobile network takes more than 2 minutes, which is unbearably long
Change-Id: Ic3e1e0b87b874fe34e2a4e83bd4cd1640234f456
Signed-off-by: BrainKub <fyhtjnt@gmail.com>
- Update rootdir chowns
- Uppercase the device name again
Co-authored-by: aswin7469 <aswinas@pixysos.com>
Change-Id: I96d837889cdbd4cd9652a7f5d5e716c3767c8aca
Moving init.qcom.testscripts.sh service start from
init.qcom.rc to init.test.rc to prevent it from having
any signature in user builds.
Change-Id: Ifa69537f8394e0919d26d6db300910620278ed76
Apart from useless battery capacity faking, all this does is to
start/stop data and radio services based upon the baseband type
of several QCOM targets. Since all the supported targets in this
common tree are of the same type, simply start all relevant services
unconditionally and get rid of one more shell script.
Change-Id: I1297b2537fcbea31d514bab5e248691cabdf523c
time_daemon starts from its own init rc
init: processing action (ro.crypto.state=encrypted && ro.crypto.type=file && post-fs-data) from (/vendor/etc/init/init.time_daemon.rc:7)
init: starting service 'time_daemon'...
Fixes:
ignored duplicate definition of service 'time_daemon'
Change-Id: Ia92d88eb2ec9696bc8ce075e74833cde572cb1e1
* W libc : Unable to set property "ctl.stop" to "ril-daemon": error code: 0x20
* W libc : Unable to set property "ctl.stop" to "vendor.ril-daemon": error code: 0x20
Change-Id: I80784cb3d6ee3cf539b3a8b5dbcada2e8b850641
