Messages in this thread | | | Date | Tue, 27 Jan 2009 17:05:36 +0100 | From | Damien Wyart <> | Subject | Re: Strange load average and ksoftirqd behavior with 2.6.29-rc2-git1 |
| |
Hello,
Just a heads up on this problem I sent a few days ago:
> Testing 2.6.29-rc2-git1, I notice something which doesn't happen with > 2.6.28.1 on the same machine with the same config (minus new config > entries for 2.6.29) : when the machine is idle, the load average > reported by top is not 0.00 (it is with 2.6.28.1) but between 0.30 and > 0.80 and either ksoftirqd/0 or ksoftirqd/1 is listed as number one in > top output, with a few % of CPU. The machine seems to work fine but > I think something is still wrong.
> Here is an example of top output :
> top - 17:23:40 up 19 min, 2 users, load average: 0.42, 0.48, 0.34 > Tasks: 99 total, 1 running, 98 sleeping, 0 stopped, 0 zombie > Cpu(s): 0.2%us, 0.5%sy, 0.0%ni, 99.1%id, 0.0%wa, 0.0%hi, 0.2%si, 0.0%st > Mem: 2075420k total, 311920k used, 1763500k free, 4912k buffers > Swap: 1212896k total, 0k used, 1212896k free, 115648k cached
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 4 root 15 -5 0 0 0 S 1 0.0 0:30.74 ksoftirqd/0 > 1 root 20 0 2100 692 588 S 0 0.0 0:00.47 init
> I attach dmesg, config and /proc/interrupts. Feel free to ask more info > if needed.
The problem is still present in 2.6.29-rc2-git3 on this P4 machine, but I do not see it on a Centrino laptop.
Config has not changed and I do not know if /proc/interrupts is useful or not so I am not attaching anything for now, feel free to ask for more information and I will do my best to provide it. I know someone else also experienced this problem on his machines.
-- Damien Wyart
| |