idle process keeping cpu 150% busy in freebsd 9.1-amd64 [solved]

Mon Jun 3 14:15:17 UTC 2013

Thanks very much to all for your help.

I finally resolved the problem: first, upon logging in, I 
changed the window system to "fluxbox", instead of my usual 
Gnome.  The cpu quieted down.  This suggested that I had 
messed up something having to do with Gnome. So I
adopted the trivial fix: I had done little work on the 
system, so I re-installed PC-BSD 9.1.  Now I am running Gnome
and both cores are fine.

One small issue remains: the system doesn't suspend 
properly. If I suspend it from the Gnome System -> "Shut 
down..." menu,  it appears to suspend, but the fans keep 
running, and it doesn't want to wake up again, even if I 
power it off.  The only way to "wake it up" is pull the 
power cord and plug it in again, and then it reboots. 
Perhaps this is a known ACPI problem?

Excerpt from dmesg:

aesni0: No AESNI support.
acpi0: <HPQOEM SLIC-CPC> on motherboard
acpi0: Power Button (fixed)
ACPI Error: Field [ASSM] at 524320 exceeds Buffer [BUF0] 
size 880 (bits) (20110527/dsopcode-254)
ACPI Error: Method parse/execution failed [\\_SB_.MEM_._CRS] 
(Node 0xfffffe0003cfc380), AE_AML_BUFFER_LIMIT 
(20110527/psparse-560)
ACPI Error: Method execution failed [\\_SB_.MEM_._CRS] (Node 
0xfffffe0003cfc380), AE_AML_BUFFER_LIMIT (20110527/uteval-113)
can't fetch resources for \\_SB_.MEM_ - AE_AML_BUFFER_LIMIT
cpu0: <ACPI CPU> on acpi0
cpu1: <ACPI CPU> on acpi0

                                                         Kostas

>
> ----- Original Message ----- From: "Kostas Oikonomou" 
> <k.oikonomou at att.net>
> To: "John Baldwin" <jhb at freebsd.org>
> Cc: <freebsd-amd64 at freebsd.org>
> Sent: Friday, May 31, 2013 9:49 PM
> Subject: Re: idle process keeping cpu 150% busy in freebsd 
> 9.1-amd64
>
>
>> The core will always look like it is "running" in top, 
>> even when it is asleep. That is just how FreeBSD accounts 
>> for idle CPU time. The only thing I was hoping would 
>> change is the fan having to run. You can try kldload'ing 
>> coretemp and seeing if the processor temperatures are 
>> different when deeper CX states are enabled (or when 
>> powerd is running) to see if it is having any affect on 
>> the temperatures in your box.
>>
>> First the good news.  It looks like the problem is solved 
>> on the laptop (Core i7). It took one more reboot after I 
>> put "performance_cx_lowest=LOW" in /etc/rc.conf.
>>
>> However, the problem is still there on the HP desktop 
>> (AMD 7550).  This has only
>> Cx state, C1, so "performance_cx_lowest=LOW" had no effect.
>>
>> The symptoms with this machine are that top does not show 
>> anything running besides idle, and neither does ps -aux.  
>> Yet the Gnome System monitor applet that I have on the 
>> bottom panel shows significant cpu activity.
>> And the fan starts running within 5 minutes after the 
>> system finishes booting.
>>
>> Here is what top -S -H says:
>>
>>
>> last pid:  2645;  load averages:  1.14,  0.78, 
>> 0.34               up 0+00:02:17  19:31:35
>> 356 processes: 3 running, 338 sleeping, 15 waiting
>> CPU:  0.2% user,  0.0% nice, 18.9% system,  0.0% 
>> interrupt, 80.9% idle
>> Mem: 187M Active, 36M Inact, 354M Wired, 13M Cache, 3323M 
>> Free
>> Swap: 2048M Total, 2048M Free
>>
>>   PID USERNAME   PRI NICE   SIZE    RES STATE   C   
>> TIME   WCPU COMMAND
>>    11 root       155 ki31     0K    32K CPU0    0   1:45 
>> 89.99% idle{idle: cpu0}
>>    11 root       155 ki31     0K    32K RUN     1   1:40 
>> 83.98% idle{idle: cpu1}
>>     0 root       -16    0     0K  2672K sched   0   1:03  
>> 0.00% kernel{swapper}
>>   462 root       -21  r31   912M 33216K select  0   0:10  
>> 0.00% Xorg
>>  1968 ko          52    0   209M  7144K select  0   0:04  
>> 0.00% pulseaudio{pulseaudio}
>>  1968 ko          52    0   209M  7144K select  1   0:03  
>> 0.00% pulseaudio{pulseaudio}
>>     7 root       -16    -     0K    16K ccb_sc  0   0:02  
>> 0.00% xpt_thrd
>>    12 root       -84    -     0K   240K WAIT    1   0:01  
>> 0.00% intr{irq1: atkbd0}
>>    12 root       -60    -     0K   240K WAIT    0   0:01  
>> 0.00% intr{swi4: clock}
>>  1969 ko          20    0   323M 21968K select  0   0:00  
>> 0.00% gnome-panel{gnome-panel}
>>    12 root       -96    -     0K   240K WAIT    1   0:00  
>> 0.00% intr{irq16: vgapci0+}
>>  2196 ko          20    0   294M 18052K select  0   0:00  
>> 0.00% gnome-netstatus-app{gnome-
>>  1811 ko          20    0   320M 19116K select  1   0:00  
>> 0.00% gnome-settings-daem{gnome-
>>    15 root       -68    -     0K   128K -       1   0:00  
>> 0.00% usb{usbus0}
>>  2200 ko          20    0   360M 21808K select  1   0:00  
>> 0.00% clock-applet{clock-applet}
>>  1458 root        30   10 10376K  3448K select  0   0:00  
>> 0.00% devd
>>  2028 ko          20    0   218M 25652K select  0   0:00  
>> 0.00% python
>>  2272 ko          20    0   280M 20044K select  1   0:00  
>> 0.00% gnome-terminal{gnome-termi
>>  2198 ko          20    0   295M 20552K select  1   0:00  
>> 0.00% stickynotes_applet{stickyn
>>  1405 ko          20    0   156M 13152K select  0   0:00  
>> 0.00% gnome-session{gnome-sessio
>>   417 haldaemon   20    0 56952K  6136K select  0   0:00  
>> 0.00% hald{hald}
>
>
> Assuming this is a dual core machine, your missing ~25% of 
> your overall
> CPU time, identifying where this is might help.
>
> Does the following give any hints to this: top -S -H -I -C 
> -s1
>
>    Regards
>    Steve
>
> ================================================
> This e.mail is private and confidential between Multiplay 
> (UK) Ltd. and the person or entity to whom it is 
> addressed. In the event of misdirection, the recipient is 
> prohibited from using, copying, printing or otherwise 
> disseminating it or any information contained in it.
> In the event of misdirection, illegible or incomplete 
> transmission please telephone +44 845 868 1337
> or return the E.mail to postmaster at multiplay.co.uk.
>