kapm-idled using lots of CPU
James R. Van Zandt
When nothing else is going on, kapm-idled is using almost all of the
CPU on my new Gateway 3350 Solo:
15:16:33 up 4:00, 2 users, load average: 0.00, 0.01, 0.00
33 processes: 31 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 0.2% user, 96.4% system, 0.0% nice, 3.4% idle
Mem: 255444K total, 161364K used, 94080K free, 23004K buffers
Swap: 248996K total, 0K used, 248996K free, 75212K cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
3 root 20 0 0 0 0 SW 96.4 0.0 230:18 kapm-idled
9096 jrv 10 0 984 984 784 R 0.1 0.3 0:00 top
1 root 8 0 532 532 464 S 0.0 0.2 0:03 init
2 root 9 0 0 0 0 SW 0.0 0.0 0:00 keventd
4 root 19 19 0 0 0 SWN 0.0 0.0 0:00 ksoftirqd_CPU0
5 root 9 0 0 0 0 SW 0.0 0.0 0:00 kswapd
6 root 9 0 0 0 0 SW 0.0 0.0 0:00 kreclaimd
Is this normal? Is it wasting energy and generating heat as I suspect?
FWIW I am running Linux 2.4.7, with these configuration parameters:
CONFIG_APM=y
# CONFIG_APM_IGNORE_USER_SUSPEND is not set
CONFIG_APM_DO_ENABLE=y
CONFIG_APM_CPU_IDLE=y
CONFIG_APM_DISPLAY_BLANK=y
# CONFIG_APM_RTC_IS_GMT is not set
# CONFIG_APM_ALLOW_INTS is not set
# CONFIG_APM_REAL_MODE_POWER_OFF is not set
I'd appreciate any pointers.
- Jim Van Zandt
Uwe Bonnes
: When nothing else is going on, kapm-idled is using almost all of the
: CPU on my new Gateway 3350 Solo:
When do you expect an idle job to take processor cycles?
--
Uwe Bonnes b...@elektron.ikp.physik.tu-darmstadt.de
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------
alexande...@ic.ac.uk
>
> When nothing else is going on, kapm-idled is using almost all of the
> CPU on my new Gateway 3350 Solo:
>
> 15:16:33 up 4:00, 2 users, load average: 0.00, 0.01, 0.00
> 33 processes: 31 sleeping, 2 running, 0 zombie, 0 stopped
> CPU states: 0.2% user, 96.4% system, 0.0% nice, 3.4% idle
> Mem: 255444K total, 161364K used, 94080K free, 23004K buffers
> Swap: 248996K total, 0K used, 248996K free, 75212K cached
>
> PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
> 3 root 20 0 0 0 0 SW 96.4 0.0 230:18 kapm-idled
> 9096 jrv 10 0 984 984 784 R 0.1 0.3 0:00 top
> 1 root 8 0 532 532 464 S 0.0 0.2 0:03 init
> 2 root 9 0 0 0 0 SW 0.0 0.0 0:00 keventd
> 4 root 19 19 0 0 0 SWN 0.0 0.0 0:00 ksoftirqd_CPU0
> 5 root 9 0 0 0 0 SW 0.0 0.0 0:00 kswapd
> 6 root 9 0 0 0 0 SW 0.0 0.0 0:00 kreclaimd
>
>
> Is this normal? Is it wasting energy and generating heat as I suspect?
>
regardless of what *any* top like proggy tells you. The reason that these
power saving instructions are now called by a kernel process (aka kapm-idled)
is because buggy BIOS can cause linux to crash. Linus decided that making
these idle calls a process only that process will bomb out rather than the
whole kernel :) I also think this may be more efficient however I could be
wrong :)
as a conclusion, this is normal, you will notice that your laptop is still
cool and not using its fan as much as windows :)
have fun
Alex
Teemu Kilpivuori
>> Mem: 255444K total, 161364K used, 94080K free, 23004K buffers
>> Swap: 248996K total, 0K used, 248996K free, 75212K cached
>>
>> PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
>> 3 root 20 0 0 0 0 SW 96.4 0.0 230:18 kapm-idled
>>
>>
> kthis has been a 'problem' from day one. Your cput is actually being idle
> regardless of what *any* top like proggy tells you. The reason that these
> power saving instructions are now called by a kernel process (aka kapm-idled)
> is because buggy BIOS can cause linux to crash. Linus decided that making
> these idle calls a process only that process will bomb out rather than the
> whole kernel :) I also think this may be more efficient however I could be
> wrong :)
> as a conclusion, this is normal, you will notice that your laptop is still
> cool and not using its fan as much as windows :)
For me, on laptops it seems to work ok, but on all desktops this thing
seems to be broken. For no reason idled starts hogging cpu, and yes,
it is NOT idle. It shows clearly when watching CPU-temperature.
Temperature goes up by 20 degrees celcius when this happens.
So I used "apm=off" in lilo.conf append line. Now idled process is gone
and looking at cpu-temperatures verify that machine is really idle.
I don't know (or care) whats broken but something definitely is broken.
On laptops (IBM Thinkpad, SNI Scenic mobile) it seems to work OK.
This happened when I was running 2.4.3. I haven't tried it with newer
kernels. I don't want apm anyway (except in laptops...)
So to verify if machine really is idle, watch temperature/fan
activity/battery life.
--
// Teemu
H. Dziardziel
<teemu.ki...@wd.utu.nospam.fi> wrote:
>
>For me, on laptops it seems to work ok, but on all desktops this thing
>seems to be broken. For no reason idled starts hogging cpu, and yes,
>it is NOT idle. It shows clearly when watching CPU-temperature.
>Temperature goes up by 20 degrees celcius when this happens.
>
>So I used "apm=off" in lilo.conf append line. Now idled process is gone
>and looking at cpu-temperatures verify that machine is really idle.
>
>I don't know (or care) whats broken but something definitely is broken.
>On laptops (IBM Thinkpad, SNI Scenic mobile) it seems to work OK.
>
>This happened when I was running 2.4.3. I haven't tried it with newer
>kernels. I don't want apm anyway (except in laptops...)
>
>So to verify if machine really is idle, watch temperature/fan
>activity/battery life.
>
>--
>// Teemu
The hlt instruction does work or it would not be part of the kernel.
However the cpu type and bios determine what happens actually.
It could be the cpu fan is turned off by the bios when apm runs,
So, the cpu temperature alone is not a reliable indicator that hlt is
working unless the fan is running/or not, identically.
Apm and hlt are two different things. Application problems ( cpu's
can go into loops and so negate hlt due to mismatched libraries as
an example) and network mantainence will keep the cpu busy etc. Any
such none app caused matters are surely bios apm implemenation caused,
not by hlt.