| commit | 23ae9c8bec4afe6ddd90233cb24c8747d0e29b03 | [log] [tgz] |
|---|---|---|
| author | Hao Jiang <jianghao@google.com> | Tue Oct 07 18:52:04 2025 +0000 |
| committer | Hao Jiang <jianghao@google.com> | Thu Oct 16 09:56:37 2025 -0700 |
| tree | 21a63c8baad2c7713bc17cdc117d7f34dd911b4f | |
| parent | 8342a7d14decaf4024efd7596aeacb1dd747e47d [diff] |
use-reset: listen to the chassis signal instead of host Inherited from the dbus-sensors, the `powerMatch` for PowerState::on was incorrectly pointed to the Host status. That is a mismatch for the definition of x86-power-control: https://github.com/openbmc/x86-power-control/blob/master/src/power_control.cpp#L515-L537 More specifically, `PowerState::checkForWarmReset` should be considered as part of PowerState::On, instead of PowerState::Off. The ChassisState of x86-power-control is more close to reflect the desired hardware signal (PowerOk signal). https://github.com/openbmc/x86-power-control/blob/master/src/power_control.cpp#L538-L559 The only delta is the `waitForSIOPowerGood` which is still considered as Chassis.PowerState.Off, which happens after the `Event::powerOKAssert`. https://github.com/openbmc/x86-power-control/blob/master/src/power_control.cpp#L1781 However, the SIO is not enabled on the target platform Google-Bug-Id: 316465377 Tested: tested against a warmreboot machine and confirmed no usb reset Change-Id: If441bdfb79f5262bbbc60235f3447ab3e0aae816 Signed-off-by: Hao Jiang <jianghao@google.com>