[email protected] (Swift Griggs) writes:
>Folks, I recently installed NetBSD on a Lenovo M83 Tiny machine and from
>time to time, I notice the "[system]" (appears to be a kernel thread?)
>getting up to 80% of the CPU while the box is doing .... nothing.
[system] is all the kernel threads. In top you can switch to thread
display and get more details. Kernel threads are also displayd with 'ps -s'
and you can augment the display with the thread name using '-o lname'.
80% CPU for doing nothing however is bad. The top display probably tells
you which thread is misbehaving.
>On another system, I have a question about a 1.8Ghz CoreDuo based 32-bit
>i386 laptop with 2GB of RAM. I noticed that '[system]' accumulates the
>most time on the host, but it's never "on the board" when I run top or
>other tools.
The kernel threads are only active for very short time intervals
and therefore appear rarely as active. But since they run all the
time, their usage accumulates. E.g.
a system with 59 days uptime:
0 ? OKl 678:30.07 [system]
that's less than 1% overall. Here are the kernel threads reports by 'ps':
% ps asx -o uid,pid,ppid,cpu,pri,nice,vsz,rss,wchan,stat,tty,time,comm,lname |
grep system
0 0 0 0 123 0 0 7728 - OKl ? 678:30.22 [system
(zombie)
0 0 0 0 96 0 0 7728 vndbp OKl ? 678:30.22 [system
vnd0
0 0 0 0 96 0 0 7728 nfsiod OKl ? 678:30.22 [system
nfsio
0 0 0 0 96 0 0 7728 nfsiod OKl ? 678:30.22 [system
nfsio
0 0 0 0 96 0 0 7728 nfsiod OKl ? 678:30.22 [system
nfsio
0 0 0 0 96 0 0 7728 nfsiod OKl ? 678:30.22 [system
nfsio
0 0 0 0 123 0 0 7728 physiod OKl ? 678:30.22 [system
physiod
0 0 0 0 96 0 0 7728 lnxcmplt OKl ? 678:30.22 [system
VCHIQka-0
0 0 0 0 125 0 0 7728 aiodoned OKl ? 678:30.22 [system
aiodoned
0 0 0 0 124 0 0 7728 syncer OKl ? 678:30.22 [system
ioflush
0 0 0 0 126 0 0 7728 pgdaemon OKl ? 678:30.22 [system
pgdaemon
0 0 0 0 123 0 0 7728 data OKl ? 678:30.22 [system
vcaudio
0 0 0 0 96 0 0 7728 semacv OKl ? 678:30.22 [system
VCHIQs-0
0 0 0 0 96 0 0 7728 semacv OKl ? 678:30.22 [system
VCHIQr-0
0 0 0 0 96 0 0 7728 semacv OKl ? 678:30.22 [system
VCHIQ-0
0 0 0 0 96 0 0 7728 cleanup OKl ? 678:30.22 [system
iscsi_cleanup
0 0 0 0 96 0 0 7728 sccomp OKl ? 678:30.22 [system
scsibus0
0 0 0 0 96 0 0 7728 usbevt OKl ? 678:30.22 [system
usb0
0 0 0 0 96 0 0 7728 npfgccv OKl ? 678:30.22 [system
npfgc
0 0 0 0 96 0 0 7728 unpgc OKl ? 678:30.22 [system
unpgc
0 0 0 0 222 0 0 7728 nd6_time OKl ? 678:30.22 [system
nd6_timer
0 0 0 0 222 0 0 7728 rt_timer OKl ? 678:30.22 [system
rt_timer
0 0 0 0 43 0 0 7728 vcmbox0 OKl ? 678:30.22 [system
vcmbox0
0 0 0 0 125 0 0 7728 vmem_reh OKl ? 678:30.22 [system
vmem_rehash
0 0 0 0 123 0 0 7728 mmctaskq OKl ? 678:30.22 [system
sdmmc0
0 0 0 0 96 0 0 7728 usbtsk OKl ? 678:30.22 [system
usbtask-dr
0 0 0 0 96 0 0 7728 usbtsk OKl ? 678:30.22 [system
usbtask-hc
0 0 0 0 43 0 0 7728 dwc2 OKl ? 678:30.22 [system
dwc2
0 0 0 0 96 0 0 7728 iicintr OKl ? 678:30.22 [system
iic0
0 0 0 0 127 0 0 7728 xcall OKl ? 678:30.22 [system
xcall/3
0 0 0 0 223 0 0 7728 - OKl ? 678:30.22 [system
softser/3
0 0 0 0 220 0 0 7728 - OKl ? 678:30.22 [system
softclk/3
0 0 0 0 221 0 0 7728 - OKl ? 678:30.22 [system
softbio/3
0 0 0 0 222 0 0 7728 - OKl ? 678:30.22 [system
softnet/3
0 0 0 0 0 0 0 7728 - OKl ? 678:30.22 [system
idle/3
0 0 0 0 127 0 0 7728 xcall OKl ? 678:30.22 [system
xcall/2
0 0 0 0 223 0 0 7728 - OKl ? 678:30.22 [system
softser/2
0 0 0 0 220 0 0 7728 - OKl ? 678:30.22 [system
softclk/2
0 0 0 0 221 0 0 7728 - OKl ? 678:30.22 [system
softbio/2
0 0 0 0 222 0 0 7728 - OKl ? 678:30.22 [system
softnet/2
0 0 0 0 0 0 0 7728 - OKl ? 678:30.22 [system
idle/2
0 0 0 0 127 0 0 7728 xcall OKl ? 678:30.22 [system
xcall/1
0 0 0 0 223 0 0 7728 - OKl ? 678:30.22 [system
softser/1
0 0 0 0 220 0 0 7728 - OKl ? 678:30.22 [system
softclk/1
0 0 0 0 221 0 0 7728 - OKl ? 678:30.22 [system
softbio/1
0 0 0 0 222 0 0 7728 - OKl ? 678:30.22 [system
softnet/1
0 0 0 0 0 0 0 7728 - OKl ? 678:30.22 [system
idle/1
0 0 0 0 96 0 0 7728 smtaskq OKl ? 678:30.22 [system
sysmon
0 0 0 0 96 0 0 7728 crypto_w OKl ? 678:30.22 [system
cryptoret
0 0 0 0 43 0 0 7728 pmfsuspe OKl ? 678:30.22 [system
pmfsuspend
0 0 0 0 43 0 0 7728 pmfevent OKl ? 678:30.22 [system
pmfevent
0 0 0 0 96 0 0 7728 sopendfr OKl ? 678:30.22 [system
sopendfree
0 0 0 0 43 0 0 7728 nfssilly OKl ? 678:30.22 [system
nfssilly
0 0 0 0 125 0 0 7728 cachegc OKl ? 678:30.22 [system
cachegc
0 0 0 0 125 0 0 7728 vrele OKl ? 678:30.22 [system
vrele
0 0 0 0 125 0 0 7728 vdrain OKl ? 678:30.22 [system
vdrain
0 0 0 0 125 0 0 7728 mod_unld OKl ? 678:30.22 [system
modunload
0 0 0 0 127 0 0 7728 xcall OKl ? 678:30.22 [system
xcall/0
0 0 0 0 223 0 0 7728 - OKl ? 678:30.22 [system
softser/0
0 0 0 0 220 0 0 7728 - OKl ? 678:30.22 [system
softclk/0
0 0 0 0 221 0 0 7728 - OKl ? 678:30.22 [system
softbio/0
0 0 0 0 222 0 0 7728 - OKl ? 678:30.22 [system
softnet/0
0 0 0 0 0 0 0 7728 - OKl ? 678:30.22 [system
idle/0
0 0 0 0 125 0 0 7728 uvm OKl ? 678:30.22 [system
swapper
CPU accounting is reported only per-process, so you can't see what thread
accumulated
how much time. top in thread mode on the other hand can give you details:
PID LID USERNAME PRI STATE TIME WCPU CPU NAME COMMAND
0 51 root 222 rt_tim/2 0:59 0.00% 0.00% rt_timer [system]
0 50 root 43 vcmbox/3 0:14 0.00% 0.00% vcmbox0 [system]
0 49 root 125 vmem_r/2 0:08 0.00% 0.00% vmem_reha [system]
0 48 root 123 mmctas/3 0:00 0.00% 0.00% sdmmc0 [system]
0 39 root 96 usbtsk/1 47:10 0.00% 0.00% usbtask-d [system]
0 38 root 96 usbtsk/0 0:00 0.00% 0.00% usbtask-h [system]
0 37 root 43 dwc2/2 0:00 0.00% 0.00% dwc2 [system]
0 36 root 96 iicint/3 0:55 0.00% 0.00% iic0 [system]
0 35 root 127 xcall/3 0:00 0.00% 0.00% xcall/3 [system]
0 34 root 223 IDLE/3 0:00 0.00% 0.00% softser/3 [system]
0 33 root 220 IDLE/3 0:00 0.00% 0.00% softclk/3 [system]
0 32 root 221 IDLE/3 0:00 0.00% 0.00% softbio/3 [system]
0 31 root 222 IDLE/3 0:00 0.00% 0.00% softnet/3 [system]
0 30 root 0 IDLE/3 0:00 0.00% 0.00% idle/3 [system]
0 29 root 127 xcall/2 0:00 0.00% 0.00% xcall/2 [system]
0 28 root 223 IDLE/2 0:00 0.00% 0.00% softser/2 [system]
0 27 root 220 IDLE/2 0:00 0.00% 0.00% softclk/2 [system]
0 26 root 221 IDLE/2 0:00 0.00% 0.00% softbio/2 [system]
0 25 root 222 IDLE/2 0:00 0.00% 0.00% softnet/2 [system]
0 24 root 0 CPU/2 0:00 0.00% 0.00% idle/2 [system]
0 23 root 127 xcall/1 0:00 0.00% 0.00% xcall/1 [system]
0 22 root 223 IDLE/1 0:00 0.00% 0.00% softser/1 [system]
0 21 root 220 IDLE/1 0:00 0.00% 0.00% softclk/1 [system]
0 20 root 221 IDLE/1 0:00 0.00% 0.00% softbio/1 [system]
0 19 root 222 IDLE/1 0:00 0.00% 0.00% softnet/1 [system]
0 18 root 0 CPU/1 0:00 0.00% 0.00% idle/1 [system]
0 17 root 96 smtask/0 0:00 0.00% 0.00% sysmon [system]
0 16 root 96 crypto/0 0:00 0.00% 0.00% cryptoret [system]
0 15 root 43 pmfsus/3 0:00 0.00% 0.00% pmfsuspen [system]
0 14 root 43 pmfeve/1 0:00 0.00% 0.00% pmfevent [system]
0 13 root 96 sopend/0 0:00 0.00% 0.00% sopendfre [system]
0 12 root 43 nfssil/1 0:00 0.00% 0.00% nfssilly [system]
0 11 root 125 cacheg/0 2:12 0.00% 0.00% cachegc [system]
0 10 root 125 vrele/0 0:54 0.00% 0.00% vrele [system]
0 9 root 125 vdrain/2 47:43 0.00% 0.00% vdrain [system]
0 8 root 125 mod_un/3 0:00 0.00% 0.00% modunload [system]
0 7 root 127 xcall/0 0:00 0.00% 0.00% xcall/0 [system]
0 6 root 223 IDLE/0 0:00 0.00% 0.00% softser/0 [system]
0 5 root 220 IDLE/0 0:00 0.00% 0.00% softclk/0 [system]
0 4 root 221 IDLE/0 0:00 0.00% 0.00% softbio/0 [system]
0 3 root 222 IDLE/0 3:34 0.00% 0.00% softnet/0 [system]
0 2 root 0 CPU/0 0:00 0.00% 0.00% idle/0 [system]
0 1 root 125 uvm/0 17:52 0.00% 0.00% swapper [system]
0 106 root 96 vndbp/3 0:03 0.00% 0.00% vnd0 [system]
0 74 root 96 nfsiod/1 0:00 0.00% 0.00% nfsio [system]
0 73 root 96 nfsiod/1 0:00 0.00% 0.00% nfsio [system]
0 72 root 96 nfsiod/0 0:00 0.00% 0.00% nfsio [system]
0 71 root 96 nfsiod/2 0:00 0.00% 0.00% nfsio [system]
0 70 root 123 physio/2 0:48 0.00% 0.00% physiod [system]
0 69 root 96 lnxcmp/0 0:00 0.00% 0.00% VCHIQka-0 [system]
0 68 root 125 aiodon/0 5:16 0.00% 0.00% aiodoned [system]
0 67 root 124 syncer/0 535:09 0.00% 0.00% ioflush [system]
0 66 root 126 pgdaem/1 8:28 0.00% 0.00% pgdaemon [system]
0 65 root 123 data/3 0:00 0.00% 0.00% vcaudio [system]
0 64 root 96 semacv/3 0:00 0.00% 0.00% VCHIQs-0 [system]
0 63 root 96 semacv/0 0:00 0.00% 0.00% VCHIQr-0 [system]
0 62 root 96 semacv/1 0:00 0.00% 0.00% VCHIQ-0 [system]
0 59 root 96 cleanu/3 0:00 0.00% 0.00% iscsi_cle [system]
0 57 root 96 sccomp/3 0:00 0.00% 0.00% scsibus0 [system]
0 56 root 96 usbevt/2 0:01 0.00% 0.00% usb0 [system]
0 54 root 96 npfgcc/3 0:16 0.00% 0.00% npfgc [system]
0 53 root 96 unpgc/3 0:00 0.00% 0.00% unpgc [system]
Most time is spent in ioflush (searching dirty pages to write them to disk),
vdrain (VFS garbage collection), and the USB driver (this is dwctwo, others
are more efficient).
Saying this, if you run a kernel with LOCKDEBUG on a system with lots
of memory, this adds a ton of overhead to the ioflush function and then
it's not impossible to see a continous 80% CPU usage for '[system]'.
But that doesn't happen with normal kernels.
>The corollary is, does NetBSD do the same work but just mask the CPU
>usage?
top just can't display CPU usage correctly for processes that are active
for very short intervals, wether kernel threads or not.
Greetings,
--
--
Michael van Elst
Internet: [email protected]
"A potential Snark may lurk in every tree."
