tree bd6e4e0b69ea6b6ff5436cadbef51f20060d03ea
parent 548d1a27aa399392c00f9ad27c8990ecd047ca34
author Thang Tran <thuutran@amperecomputing.com> 1663860049 +0700
committer Thang Tran <thuutran@amperecomputing.com> 1664246479 +0000

replace static converting with rounding

Issue: When the users request to get the sensor thresholds via "ipmitool
sensor get <sensor_name>" command, sometimes the value is not correct.
E.g: The real sensor threshold is 2.01, the "scale" = 0, "exponentR"=-3,
"coefficientM" = 10. After calculating, the ipmid should return 201 to
ipmitool, but the actual return value is 200. Ipmitool displays the
value of the threshold is 2.00.

Root cause: The "float point" can not present exactly the "double" value
in somecase. Therefore, after calculating, the value of sensor
thresholds is incorrect when convert from double to uint8_t.

Solution: Replace static converting with rounding.

Tested:
   1. Check the result of the threshold whose value is 2.01
      ipmitool sensor get <2.01_Sensor_name>
   2. The threshold value is 2.01

Signed-off-by: Thang Tran <thuutran@amperecomputing.com>
Change-Id: I7f3e72640beb76eda9370f60e32f42d2b2c1fbda