RE: Status of ACPI on IBM T40p & D600

-----Original Message-----
From: Alexander Gran [mailto:alex-4mRm4CWMOdkHuPwBo6/EyUB+6BGkLq7r@public.gmane.org]

Am Dienstag, 10. Februar 2004 03:07 schrieben Sie Ow Mun Heng

quoted

Mine shows something like this
 C1:      promotion[C2] demotion[--] latency[000] usage[00022310]
*C2:      promotion[C3] demotion[C1] latency[050] usage[00248795]
 C3:      promotion[--] demotion[C2] latency[050] usage[00000090]

Seems like someone is still using the busmaster feature..
Perhaps some other driver, or load (Note that e.g. harddisk 
access, graphic io 
can also trigger the C2 state)

I did a few more experiments last night..
I booted into the usb-aware kernel and found that if I removed
usb-uhci module, bus-master activity will not show fffff anymore and
in virtual console (GDM running on console F7) I will see C3 increasing
with some variation in the bus-master activity [eg: 123aas134 sd2341 etc..]

however, if I then proceeded to login, I note that bus-master starts
to show fffff again until the HD access lights turn off and then it
returns to non fffff status (Again, this is seen under virtual console)
XMMS/glxgears being loaded will also see the same result.

I also noted that when logged into X (Gnome), Xterm terminal
bus-master activity will be as23asd33 etc.. but C3 will never increase
{never as in maybe just 1 or 2 counts after X mins]

If I were to switch into virtual console, then C3 will start to move again.
I've tried removing nearly every module I see in lsmod but C3 won't increase
under X. There was 1 and only 1 time during my test last night that I saw 
C3 increasing under X. But I can't replicate that again. It only happened
once!!!

X is doing bus-mastering??
 

quoted

battery life  = 2h26m13s	ave power drain = 15647mW
battery life2 = 2h28m35s	ave power drain = 15391mW

That's not much, indeed.

I know and that;s why it's a pain to live with that kind of
batt life. Anyone with a D600 can tell me if my laptop's 
or laptop's battery scrrewy?? I'll like to throw it back to DELL.

Hmm.. I wonder if anyone can post their ave power drain over
the life of the battery. I'm expecially interested in the comparison 
with T40/T41 and centrino based laptops.

it's a very simple bash script that reads /proc/acpi/battery/BAT0/state 
every 10 secs.

quoted

Another thing, should I expect the P states to change when going
from AC to Batt? (I seee it's still stucked in P2 which is 1.4Ghz
even if I run on batt.

Ouch. No, they do not switch automagically, but you should 
really switch them (either manually or via cpudynd/cpufreqd which is even 
better) to the lower values if you wan't real batterylife.

Okay.. they do not switch.. maybe that's my problem..
I'm currently using cpufreqd and cpudynd to do exactly that. but when 
I'm on battery, I will usually just disable cpudynd (as it has a tendency 
to want to switch to higher freq)

I tried echo 6>/proc/acpi/processor/CPU0/performance  but the darn thing
never budges from P2 state!!

Any idea how to reach P6 state?

Thanks for any comments/suggestions.

My Aim here is simple.. Get longer battery life!! (over the measly 2h30min 
and this is a Centrino laptop!!)



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn