Hi all,
I'm having an issue with VirtualBox where CentOS on a Windows host is spending most of the CPU time waiting for si. I've set the frequency to 100 with 'divider=10', but that seems to have no affect.
I imported the exact same virtual machine to the same version of VirtualBox running on OS X 10.7 and the guest ran normally.
I executed 'tar -cvpf test.tar.bz2 /usr/share' and while 60% of the CPU was waiting on si I did 3 polls of /proc/softirqs at approximately 2 second intervals. It looks like the vast majority of soft interrupts are for TIMER and RCU.
Is there anything else I should check before filing a bug report for this issue?
02:50:47 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %idle
02:50:48 PM all 31.94 0.00 8.33 6.94 0.00 52.78 0.00 0.00 0.00
02:50:49 PM all 34.25 0.00 0.00 0.00 0.00 65.75 0.00 0.00 0.00
02:50:50 PM all 38.03 0.00 2.82 5.63 0.00 53.52 0.00 0.00 0.00
02:50:51 PM all 40.85 0.00 0.00 0.00 0.00 59.15 0.00 0.00 0.00
02:50:52 PM all 25.64 0.00 10.26 2.56 0.00 61.54 0.00 0.00 0.00
02:50:53 PM all 37.50 0.00 1.39 0.00 1.39 59.72 0.00 0.00 0.00
02:50:54 PM all 33.33 0.00 2.78 0.00 0.00 63.89 0.00 0.00 0.00
02:50:55 PM all 35.62 0.00 6.85 5.48 0.00 52.05 0.00 0.00 0.00
Poll 1:
CPU0
HI: 0
TIMER: 213466
NET_TX: 858
NET_RX: 6201
BLOCK: 28619
BLOCK_IOPOLL: 0
TASKLET: 2
SCHED: 0
HRTIMER: 6213
RCU: 218759
Poll 2:
CPU0
HI: 0
TIMER: 214240
NET_TX: 859
NET_RX: 6216
BLOCK: 28623
BLOCK_IOPOLL: 0
TASKLET: 2
SCHED: 0
HRTIMER: 6213
RCU: 219534
Poll 3:
CPU0
HI: 0
TIMER: 215141
NET_TX: 860
NET_RX: 6228
BLOCK: 28623
BLOCK_IOPOLL: 0
TASKLET: 2
SCHED: 0
HRTIMER: 6213
RCU: 220435
Guest OS: CentOS 6.3
Guest Kernel: 2.6.32-279.19.1.el6.x86_64
Guest Additions: 4.2.6
# /etc/init.d/vboxadd status
The VirtualBox Additions are currently running.
# /etc/init.d/vboxadd-service status
Checking for VBoxService ...running
Host OS: Windows 7
VirtualBox version: 4.2.6 r82870
If necessary, I can provide the VM log.