By putting them on the bigger cores, they can be processed quicker,
resulting in improved responsiveness.
Change-Id: I13e710c94ff80cb5a044547989bda9a56d94b9e7
Signed-off-by: romgharti <romgharti@gmail.com>
* It's not using for 5.4/5.10/5.15 QCOM kernel family
Change-Id: I4d165a95f597df47f4d2ec07199b5dac0a347b61
Signed-off-by: NikitaSavc <micraftur@gmail.com>
* 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
Change-Id: I03746ea83251d8b731afb8d307526d15320c3cf9
Signed-off-by: artemscine <artempemine@gmail.com>
This CL adds build properties enabling Adaptive Suspend's short
suspend backoff with a short suspend threshold of 2 seconds and
a max backoff time of 40s. This is the same configuration that
was deployed on the previous generation of devices.
The plan is to monitor suspend behavior, screen-off uptime,
wakeup rates, and the distribution of screen-off mA on
these devices in comparison to those same attributes on late
UDC builds, which will inform whether we ship or revert.
from https://android.googlesource.com/device/google/gs201/+/refs/heads/main/device.mk#1114
Bug: 208947015
Test: booted with a test kernel and used alarmtimer mods to
exercise high-frequency wakeup behavior
Change-Id: Icf75c4dd705050833df5e48aaf0c28afa5754adb
- Add swappiness to 60
Test: boots with correct settings
Bug: 299975853
Change-Id: I623a7687ef692f8f7a5cc906e186edde294b30be
Signed-off-by: Martin Liu <liumartin@google.com>
* Because Android won't do it on a dirty flash so certain elements will be broken if something is changed in UI.
Signed-off-by: Gustavo Mendes <gusttavo.me@outlook.com>
Change-Id: I611d082319c1d90b2a133e207516d5a2e9a33512
Signed-off-by: angelomds42 <angelomds42@gmail.com>
Signed-off-by: kibria5 <mdkibria687@gmail.com>
- In my testing, this caused less jank and killed lags on
system background processes like ScreenRecorder without causing
higher power consumption.
[angelomds42: Reduce background tasks to 2-3]
Change-Id: I0068a2c1f26da8ec77f2854100a8c95b70c94be3
Signed-off-by: angelomds42 <angelomds42@gmail.com>
* The text 'TurboPower connected' will be displayed on keyguard instead of 'Charging Rapidly'
Signed-off-by: Tony Jose <gmoto1179@gmail.com>
Change-Id: I1afabade3bb7b32d8e36d521fa3d7abca9287c03
No longer needed now init listens for property changes on a
separate thread.
Bug: 186580823
Test: Cuttlefish boots successfully
Change-Id: Ief3a546d2154231d18957c2f703bd128966d7288
* Google has provisioned this on the Pixel 5 (RQ1A.201205.010) to utilize the 8 GB of RAM better, and I have also observed that a lightweight selection of apps was hitting the cached app limit before exhausting RAM on another device with 8 GB of RAM
* This shouldn't cause adverse effects on devices with less RAM (e.g. 4 GB) as the low memory killer should kick in long before this limit on such devices
Signed-off-by: Joker-V2 <nipinna0@gmail.com>
* Background ART heap compaction for cached apps has been recommended for Android Go devices for a while, and Google has enabled it for the Pixel 5 on RQ1A.201205.010. This feature should help fit more cached apps in memory
Signed-off-by: Joker-V2 <nipinna0@gmail.com>
When opening and closing activities in Settings, a measurably
significant amount of CPU time is spent processing and logging stats
events in statsd:
0.02% /apex/com.android.os.statsd/lib64/libstatspull.so @plt
0.01% /apex/com.android.os.statsd/bin/statsd @plt
0.01% /apex/com.android.os.statsd/bin/statsd SocketListener::runListener()
0.01% /apex/com.android.os.statsd/bin/statsd android::os::statsd::LogEvent::parseBuffer(unsigned char*, unsigned long)
0.01% /apex/com.android.os.statsd/bin/statsd android::os::statsd::LogEvent::parseAnnotations(unsigned char, int)
Over longer device uptimes with real-world usage, statsd uses a
substantial amount of total CPU time. We have no use for the stats
recorded by statsd, so disable the service entirely to save CPU in
potential hotpaths.
Test: simpleperf record -a; verify that statsd no longer appears in
sample hits
- Introduce a new shell script to set IRQ affinity, since different
kernels can have different IRQ numbers.
Change-Id: I61d2350e6b40f096714148c568489579bc4d2371
Co-Authored-By: Alex Naidis <alex.naidis@paranoidandroid.co>
Co-Authored-By: Jake Weinstein <jake@aospa.co>
These are critical for display performance and should not be balanced
in order to improve latency and responsiveness.
After unbalancing, move these IRQs to separate cores
to improve latency.
Set drm_vblank_offdelay parameter to -1, to make DRM
issue the vblank disable immediately with out delay.
By defailt the DRM framework has a delay of 5 secs
timeout before issuing the vblank disable.
RescueParty's aggressive approach to error recovery almost never helps
with the problems we're likely to face, and its risk of data loss is far
too high. Force-disable it entirely and remove all other conditions,
especially those that depend on data integrity in the first place, to
reduce the risk of RescueParty triggering falsely and causing data loss.
Note that it is still possible to force-enable RescueParty for testing
purposes using the persist.sys.enable_rescue property.
* Automatic Channel Selection algorithms and implementations
are used to enable interfaces to automatically figure
out which channel configuration to use for initiating
communication, for any mode of operation which initiates
radiation (AP, Mesh, IBSS, P2P).
https://wireless.wiki.kernel.org/en/users/documentation/acs
Implement a smarter system on whether to connect
to the provided 2Ghz or 5Ghz network.
Change-Id: Idd41e3d9a5e2bde3ea9b4873dffc4274d1530324
Signed-off-by: Alex Naidis <alex.naidis@linux.com>
