Update tflite models with metadata containing input data ranges.
Enable input data validation and backup sesor.
Add offsets based off temperature ranges.
Bug: 319167815
Bug: 328789394
Test: Build and boot on device. Test backup sensor and offsets
Change-Id: Icdc0c1a1fe026613e4a1716747335e60aca4edf9
Signed-off-by: Sayanna Chandula <sayanna@google.com>
- vs formulas generated from EVT 1.1 data
Bug: 320714468
Test: push to device and boot to home
Change-Id: Ife745ac7640ffd284404212706c9bdfdb0500f33
Signed-off-by: Jikai Ma <jikai@google.com>
(cherry picked from commit 859dc39412)
Inefficient OPPs were removed for the LITTLE cores from em table.
This change is to update the cdev ceilings in thermal policy to reflect
the EM changes
Bug: 329876112
Test: Build and boot on device.
Change-Id: I8b9ce12fc22dce0374d2a68a947bd598fd26f936
Signed-off-by: Ramya Subramanian <rsubr@google.com>
tz_temp callback will invoke thermal_zone_device_update and return the cached temp from thermal core. However, during device resume, the cached temp may be reset to INVALID number before being returned.
So, change the reading path to upstream temp_show which will return the temperature from thermistor driver directly.
Bug: 326252555
Test: adb shell dumpsys thermalservice
Change-Id: I82cf81d940585c4827256804bc9c9397f4b54ccf
- remove ODPM from legacy formulas
- remove ODPM from patched linear formulas
- update VIRTUAL-SKIN-CHARGE formulas
Bug: 320714468
Test: build and boot to home
Change-Id: I974476d45d9ace1740f9fec06af9399db5b7cb59
Signed-off-by: Jikai Ma <jikai@google.com>
Align with new GPU available frequency
Bug: 318313844
Test: Run gpu stress test
Change-Id: Iac169739bc72c047afa2564edb540468f48fd3ea
Signed-off-by: Peter (YM) <peterym@google.com>
- add to legacy formulas
- add to historic linear formulas
Bug: 320714468
Bug: 315234782
Bug: 307824060
Test: build and boot to home
Change-Id: If395b3281ad180731e93953089c1ca57afdb6b68
Signed-off-by: Jikai Ma <jikai@google.com>
- Updated vt predictions models with 15 samples average for ODPM rails.
- This reduces VT prediction being very sensitive to power changes
- Remove other ODPM rails from VT prediction except for VSYS_PWR_CAMERA
Bug: 307824060, 322910529
Test: Build and boot on device
Change-Id: Ibb285269d3da86e3d331ea3c7c5a3cd715da0ba2
Signed-off-by: Sayanna Chandula <sayanna@google.com>
- Switch to VIRTUAL-SKIN-MODEL from VIRTUAL-SKIN-LEGACY
- Add a backup shutdown based on VIRTUAL-SKIN-LEGACY
- Remove unnecessary polling delays from few sensors
Bug: 321108353
Test: Build and boot on device
Change-Id: I8b8909ded87ba2b76d5712c474120dbe5ab13c2c
Signed-off-by: Sayanna Chandula <sayanna@google.com>
Update thermal config as:
- Rename VIRTUAL-SKIN to VIRTUAL-SKIN-LEGACY
In order to differentiate different vt models, we need to define
legacy equation as a sepearate sensor and use it as a reference
for virtual-skin. Later when we want to switch to other models,
we could simply replace the combination for virtual-skin sensor
- Record Stats for
- VIRTUAL-SKIN-MODEL
- VIRTUAL-SKIN-MODEL - VIRTUAL-SKIN
- At intervals of 500 mdeg (-4500 to 4500)
- At intervals of 1000 mdef (-13000 to 12000 exlucding
-4000 to 4000)
- Abnormality Monitoring
- VIRTUAL-SKIN-MODEL [0-55 deg]
- VIRTUAL-SKIN-MODEL-LEGACY-DIFF [-12000 - 12000 mdeg]
Bug: 320347462
Test: Push config and verify logcat and lshal
Change-Id: I397c8a07cf01abd5d8a7dedf290829a19e21ae87
The battery_mitigation takes over the brownout event
and informs subsystems to do mitigation.
Bug: 319285945
Change-Id: I58e0a9e0469575096f34243e73c1e32474eea006
Signed-off-by: samou <samou@google.com>
This reverts commit 520b88f848.
Reason for revert: KM4 DVFS v2 tables are not merged to ToT yet. Need to keep this open until thats merged
Change-Id: Id3848230a675ef7a684f7f40a9b7083492dfab58
Bug: 318313844
Test: Tested on EVT1.0(A0) TK4 and Tskin controlled to 45C with MH offscreen, all 3 devices booted home
Change-Id: I3e10a1cd9c4493c9bb5f6a08a64e68b04713c41a
Signed-off-by: Ramya Subramanian <rsubr@google.com>
migrate from proto devices
fix power rail error in the previous update
Bug: 296476775
Bug: 315251968
Test: boot to home
Change-Id: Ie8ebfb23253f3b7c8fa2c9e723f5dd3220f55509
Signed-off-by: jikai@google.com
Bug: 310748549
Test: sideloaded the cofig files and device rebooted to home
Change-Id: Ib2af1fd19099b58fd6ba23bde10ea59767b67e05
Signed-off-by: Ramya Subramanian <rsubr@google.com>
formulas updated with P1.1 test data
Bug: 296476775
Test: boot to home
Sign-off-by: jikai@google.com
Change-Id: I88febd4792f9c32c6fcbd3d27490cdac80f00311
Removing soc tzone in json in response to non BCL loaded configuration.
Bug: 308153421
Test: Flash and boot to home
Change-Id: I4a7d247eeccf0457b5f380b7dd73058342a7f8d6
Signed-off-by: George Lee <geolee@google.com>
Sapwood configuration updates.
Bug: 308002414
Test: Flash and boot
Change-Id: I170fe1a11f8595399757f9dd5068533ef5a5da0c
Signed-off-by: George Lee <geolee@google.com>
Bug: 297083757
Test: Tested on CM4/KM4, boot to home and with thermal recorder
Change-Id: I489613010ae04851bd713d3d9c74ede15661d191
Signed-off-by: Ramya Subramanian <rsubr@google.com>
gpu cooling device is now available in kernel, enabled in HAL
Bug: 305017362
Test: tested on CM4 and KM4 devices. no access to TK4 or EVT
Change-Id: I06c022eb68a5726802529fe67fd76b189df14d32
Signed-off-by: Ramya Subramanian <rsubr@google.com>
