aocxd sets thread scheduler to SCHED_FIFO. This is so audio processing
in aocxd can run without glitching.
vndbinder:11464: type=1400 audit(0.0:17): avc: denied { sys_nice } for capability=23 scontext=u:r:aocxd:s0 tcontext=u:r:aocxd:s0 tclass=capability permissive=0
BUG: 318791959
Change-Id: I9c9148aa7b18ce525091f93956e112b4c178a129
Add new aocxd server domain
- Allow aocxd to access AOC resources
- Add new aocx binder vendor service
Allow audio hal to find and talk to aocx
avc error tcontext=u:object_r:binder_device:s0 tclass=chr_file or tcontext=u:object_r:vndbinder_device:s0 tclass=chr_file
avc: denied { add } for pid=1073 uid=0 name=aocx.IAocx scontext=u:r:aocxd:s0 tcontext=u:object_r:aocx:s0 tclass=service_manager
avc: denied { call } for scontext=u:r:hal_audio_default:s0 tcontext=u:r:aocxd:s0 tclass=binder
BUG: 315853303
Change-Id: Ide16a2be9f032bef60f43d4d3daa6372ae06b057
With multiclient HAL, the socket server domain changes from chre to
hal_contexthub_default.
Bug: 248615564
Test: updated the sepolicies and observed that avc violation logs
disappears.
Change-Id: Ic5717cee6d2714bec49814a1b779266de79dc4f3
It's possible that the gti sysfs node is not ready when
vendor.device.modules.ready is set so change the gti file owner
at later stage.
Bug: 285327052
Bug: 285247095
Bug: 285323846
Test: trigger bugreport and the dump succeed.
Change-Id: If1921aeb351603b52164c66bc9c7b3651d171287
Signed-off-by: Yen-Chao Chen <davidycchen@google.com>
Cpufreq data is sorted in ascending order, but power stats
are sorted in descending order. Reverse sorting to maintain
consistency with other power stats.
Bug: 317429309
Test: vts-tradefed run vts -m VtsHalPowerStatsTargetTest
Change-Id: I29a58c02add846eb66a8d6eb247336111363c335
Signed-off-by: Darren Hsu <darrenhsu@google.com>
Display team introduces new refresh rate residency in kernel.
Current display data provider is polling data from state sysfs.
To reduce CPU loading and improve efficiency, we should get
rid of polling data provider and create new data provider to
read data from time_in_state sysfs in PowerStatsHAL.
Bug: 316260832
Test: vts-tradefed run vts -m VtsHalPowerStatsTargetTest
Change-Id: I4d9886f13207e41f13defd89ea2c19614918a570
Signed-off-by: Darren Hsu <darrenhsu@google.com>
When rq_affinity is 1, it will default to servicing the softirq on the
same CPU that has serviced the interrupt.
By setting it to 2, completions are processed by the CPU core that
submitted the I/O.
Bug: 311655524
Signed-off-by: Qais Yousef <qyousef@google.com>
Change-Id: I21421f97f069083793fe2e2fa51739dac12eca1d
am skip reason: Merged-In I96f8d8da1169f0a428112eb27f97d215282af92e with SHA-1 27366c3935 is already in history
Original change: https://android-review.googlesource.com/c/device/google/gs-common/+/2865168
Change-Id: I828a1c722f48148c60c60d6bb92487137b0bec07
Signed-off-by: Automerger Merge Worker <android-build-automerger-merge-worker@system.gserviceaccount.com>
am skip reason: Merged-In I96f8d8da1169f0a428112eb27f97d215282af92e with SHA-1 27366c3935 is already in history
Original change: https://android-review.googlesource.com/c/device/google/gs-common/+/2865168
Change-Id: Ic0e6e87a38388730a8abdaf6fb149b67c13306f0
Signed-off-by: Automerger Merge Worker <android-build-automerger-merge-worker@system.gserviceaccount.com>
am skip reason: Merged-In I96f8d8da1169f0a428112eb27f97d215282af92e with SHA-1 27366c3935 is already in history
Original change: https://android-review.googlesource.com/c/device/google/gs-common/+/2865168
Change-Id: I82d6965f59c0caf26a4b597bd667db8516a0ff4e
Signed-off-by: Automerger Merge Worker <android-build-automerger-merge-worker@system.gserviceaccount.com>
