Age | Commit message (Collapse) | Author |
|
Add support to selectively configure thermal HAL severity for
throttling threshold for any HAL config.
Signed-off-by: Priyansh Jain <quic_priyjain@quicinc.com>
|
|
There are some thermal zones those can be created post thermal
hal initialization. It leads to a case where thermal zone
sensor init is failing due to configured sensor is not present
in available thermal zones. To avoid this issue, earlier added
a workaround to restart thermal-hal later.
thermal netlink protocol supports one event notification for
thermal zone creation. Add support to monitor thermal zone
creation event and re-initialize hal sensors.
Remove earlier workaround to restart hal service.
Change-Id: Id1aa6c3b08ada05f4c276d5b749e2c452a3fe2c6
|
|
Netlink Sample event and Violation event occurs back to back when a
threshold or hysteresis is violated or cleared. Since these events
happen in two separate threads, there is a possibility that violation
event gets executed before the sample event acquires the same mutex.
This will make the thermal severity estimation to evaluate based on the
old temperature data, since the sample event has not updated the
temperature yet.
To address the race condition, evaluate severity in the sample event
also. This will ensure that the severity is evaluated if the sample
event executes later.
Change-Id: I61fba0b07a700e0a0619d57a16d4b969e1feb57d
|
|
Add support for new thermal HAL v2 binary, which can listen to the thermal
framework netlink messages instead of the uevents sent by the userspace
governor. This new binary should be selected only for the kernel version
supporting the new netlink messages in thermal framework.
Change-Id: I31cdb93fb73d59ff3eb94eb76d6e42b3448c3ee9
|