[ltp] X60s: Heat, power management, battery

[Laurent Gilson]

Hello,

>> hdparm -B 128 /dev/your-disc.
>
> Is this safe on a SATA disk (i.e. scsi emulation)?

safe yes, successful ... maybe.

>>> - I rarely hear the fan kick in when in Linux
>>
>> That contradicts:
>>
>>> - Battery life is only about 4 hours under Linux, but upto 7 hours in  
>>> Windows.
>
> What I meant was - I don't hear it kick in (or out). Either it is on all  
> the time or it is off all the time. Under Windows, I can hear it coming  
> on or going off every now and then.

OK. That is still a bit odd. But more possible than the all the time off i  
asumed.

>> Looks like a speedstep problem.
>>
>> Are ACPI & Co loaded ?
>
> Yes.
>
>> Does /sys/devices/system/cpu/cpu0/cpufreq/ exist ?
>
> [root@vader] ~# ls -al /sys/devices/system/cpu/cpu0/cpufreq/
> total 0
> drwxr-xr-x 2 root root    0 Jul 12 15:17 ./
> drwxr-xr-x 5 root root    0 Jul 12 15:17 ../
> -r--r--r-- 1 root root 4096 Jul 12 15:17 affected_cpus
> -rw-r--r-- 1 root root    0 Jul 12 20:55 scaling_setspeed
...

looks good.

>> Does the frequency in /proc/cpuinfo change depending on the load ?
>
> Haven't checked, did so just now, and it reports
>
> vendor_id       : GenuineIntel
> cpu family      : 6
> model           : 14
> model name      : Genuine Intel(R) CPU           L2400  @ 1.66GHz
> stepping        : 8
> cpu MHz         : 1000.000
>
> on both CPUs (Core Duo) which doesnt seem right - it is powered on AC,  
> and has no reason not to be running at full 1.66 GHz.

It is running at 1GHz. The "cpu MHz" line is important for the frequency.  
"model name" is just a static string returned by the CPU. 1GHz is the  
power-save frequency. So speedstep works like it should.

Could you please run cat /proc/acpi/processor/CPU/power and post the  
output ?

I´m thinking about the USB vs CPU-sleep "bug". A centrino CPU (PentiumM,  
CoreSolo, CoreDuo,... ) can scale it´s frequency (and voltage) and go to  
sleep (C-States). The frequency-scaling is based on the load and linux can  
enforce it anytime. Sleeping is controlled by the CPU and stops working if  
something is using a lot of bus master transfers. These happen then a  
USB-device (like the internal card reader or bluetooth device) is  
connected. There are workarounds and patches against that. Newer kernels  
should not cause that.


The next point to check is the WLAN-card. iwconfig should output something  
like this:

eth0      IEEE 802.11g  ESSID:"XXXXXX"
           Mode:Managed  Frequency:2.462 GHz  Access Point:  
00:12:17:49:B9:33
           Bit Rate=54 Mb/s   Tx-Power=10 dBm
           Retry limit:7   RTS thr:off   Fragment thr:off
           Encryption key:off
           Power Management:on
           Link Quality=79/100  Signal level=-50 dBm  Noise level=-82 dBm
           Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
           Tx excessive retries:0  Invalid misc:2   Missed beacon:0

The important line is the "Power Management". off would be bad. "iwconfig  
eth0 power period 10000" works good with intel 2200 chips.


And is it a model with touchscreen ?


cu