Skip to content

top: fix cpu utilization display issue when virtual machine is present

hdzhoujie requested to merge hdzhoujie/procps:master into master

The v3.3.17 version is different from the v4.0.0 version to obtain the total time method

v3.3.17:

total = user + nice + system + idle + iowait + irq + softirq + steal

v4.0.0:

total = user + nice + system + idle + iowait + irq + softirq + steal + guest + guest_nice

When the physical machine has multiple virtual machines, the cpu utilization rate display is different, and the disp lay results of the two versions are different as follows：

version v3.3.17:

%cpu0 :100.0 us, 0.0 sy, 0.0 ni, 0.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st

version v4.0.0:

%Cpu0 : 50.1 us, 0.0 sy, 0.0 ni, 0.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st

Add the missing 'guest' tics to existing system ('sy') percentages. After modification, the result is shown as follows：

%Cpu0 : 50.1 us, 49.9 sy, 0.0 ni, 0.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st

signed-off-by:zhoujie zhoujie133@huawei.com

Edited by hdzhoujie

Merge request reports