Dell Inspiron 6000

[juniper]

amaroc: just to be clear, do you have an mplayer running on a video file at all times, or just when you enter an STR cycle - i.e. do you then close mplayer after wake up?

another thing. i have noticed that xf86-video-i810 versions 1.6.0 and 1.6.1 are in portage and in ~arch. i am running an arch system. i could unmask that and install and see how it goes, but i don't know how that will play with modular xorg. if it is like other stable/unstable packages, will it just want to emerge its dependecies? or do i have to upgrade all of xorg to 7.1? btw, i have xf86-video-i810 version 1.4.1.3.

[nybbles]

So I followed the howto on solving common ACPI problems and found that compiling my DSDT returned some warnings.

Does that mean that my DSDT is broken.. or is it only broken if there are actually errors upon recompilation? Perhaps this could be the reason for my laptop lid not registering events and always being in the 'closed' state.

[amaroc]

amaroc: just to be clear, do you have an mplayer running on a video file at all times, or just when you enter an STR cycle - i.e. do you then close mplayer after wake up?

If I ever have started xine I let one instance of xine up and end it never.

I've tested here again w/ mplayer - i can't get it crash xorg even if no video-app was running during STR. It's only xine (or xinelib-based) that forces xorg to die. That's definitely a different behaviour to your installation - don't ask me why.

another thing. i have noticed that xf86-video-i810 versions 1.6.0 and 1.6.1 are in portage and in ~arch. i am running an arch system. i could unmask that and install and see how it goes, but i don't know how that will play with modular xorg. if it is like other stable/unstable packages, will it just want to emerge its dependecies? or do i have to upgrade all of xorg to 7.1? btw, i have xf86-video-i810 version 1.4.1.3.

I'm running Xorg 7.0-r1. Xorg.log tells me the i810 driver is 1.4.1 - where did you get the 1.4.1.3 from?

As stated earlier - it might be worth to test with 7.1. I've read somewhere that there are only some dependencies - at least much less than upgrading from 6.x. Don't know if 7.0 would work w/ 1.6.x drivers.

What makes me wonder is the different behaviuor of mplayer between your machine and my. I assume you are not using KDE what could be one reason. The other one might be flags or gcc. Here is what I have:

Code: Select all

# gcc-config -c
i686-pc-linux-gnu-3.4.6

# emerge --info
Portage 2.1-r2 (default-linux/x86/2005.1, gcc-3.4.6, glibc-2.3.6-r4, 2.6.17-gentoo-r4 i686)
=================================================================
System uname: 2.6.17-gentoo-r4 i686 Intel(R) Pentium(R) M processor 1.60GHz
Gentoo Base System version 1.12.4
app-admin/eselect-compiler: [Not Present]
dev-lang/python:     2.3.5, 2.4.3-r1
dev-python/pycrypto: 2.0.1-r5
dev-util/ccache:     [Not Present]
dev-util/confcache:  [Not Present]
sys-apps/sandbox:    1.2.17
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2
sys-devel/binutils:  2.16.1-r3
sys-devel/gcc-config: 1.3.13-r3
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-O2 -march=pentium3 -pipe"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/lib/mozilla/defaults/pref /usr/share/X11/xkb /usr/share/config"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/gconf /etc/java-config/vms/ /etc/revdep-rebuild /etc/terminfo"
CXXFLAGS="-O2 -march=pentium3 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig distlocks metadata-transfer sandbox sfperms strict"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LANG="de_DE.utf8"
LINGUAS="de"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude='/distfiles' --exclude='/local' --exclude='/packages'"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 X aac alsa amarok apm arts avi berkdb bitmap-fonts cdparanoia cdr cdrom cli crypt cups dlloader dri dvd dvdr dvdread eds emboss encode firefox foomaticdb fortran gdbm gif gpm gstreamer i8x0 imagemagick imap imlib ipv6 isdnlog jpeg kde libg++ libwww mad mikmod motif mozilla mp3 mpeg ncurses nls ogg oggvorbis opengl oss pam pcre pdflib perl png pppd python qt3 qt4 quicktime readline real reflection sdl session slang spell spl ssl tcpd theora truetype truetype-fonts type1-fonts unicode usb vorbis win32codecs xml2 xmms xorg xv xvid zlib elibc_glibc input_devices_keyboard input_devices_mouse input_devices_synaptics input_devices_evdev kernel_linux linguas_de userland_GNU video_cards_i810 video_cards_vesa video_cards_fbdev video_cards_dummy video_cards_vga"
Unset:  CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LC_ALL, LDFLAGS, PORTAGE_RSYNC_EXTRA_OPTS, PORTDIR_OVERLAY

@nybbles

So I followed the howto on solving common ACPI problems and found that compiling my DSDT returned some warnings.

Does that mean that my DSDT is broken.. or is it only broken if there are actually errors upon recompilation? Perhaps this could be the reason for my laptop lid not registering events and always being in the 'closed' state.

I don't think it is needed to recompile DSDT. But because you have the ATI version I can only guess.
- check BIOS version - I'm running A08 - but A09 is the latest
- check your boot messages for ACPI&LID

Code: Select all

ACPI: AC Adapter [AC] (on-line)
ACPI: Battery Slot [BAT0] (battery present)
ACPI: Lid Switch [LID]
ACPI: Power Button (CM) [PBTN]
ACPI: Sleep Button (CM) [SBTN]
ACPI: Video Device [VID] (multi-head: yes  rom: no  post: no)
ACPI: Video Device [VID] (multi-head: yes  rom: no  post: no)
ACPI: Video Device [VID2] (multi-head: yes  rom: no  post: no)
ACPI: CPU0 (power states: C1[C1] C2[C2] C3[C3] C4[C3])
ACPI: Processor [CPU0] (supports 8 throttling states)
ACPI: Thermal Zone [THM] (45 C)
...
ACPI wakeup devices: 
LID PBTN PCI0 USB0 USB1 USB2 USB4 USB3 MODM PCIE 
ACPI: (supports S0 S3 S4 S5)

and you should see some ACPI events other then LID, e.g.

Code: Select all

logger: ACPI group ac_adapter / action ac_adapter is not defined
logger: ACPI group battery / action battery is not defined
...
video bus notify
video bus notify
logger: ACPI group video / action video is not defined
logger: ACPI group video / action video is not defined
logger: ACPI action: LID OPEN

[juniper]

I've tested here again w/ mplayer - i can't get it crash xorg even if no video-app was running during STR. It's only xine (or xinelib-based) that forces xorg to die. That's definitely a different behaviour to your installation - don't ask me why.

that is weird. mplayer definitely crashes things (all of them do). you are right, i am using gnome, not kde. i somehow doubt that is the problem, but you never know. i don't think i am going to switch for this though.

I'm running Xorg 7.0-r1. Xorg.log tells me the i810 driver is 1.4.1 - where did you get the 1.4.1.3 from?

it is in portage.

Code: Select all

eix -s i810
* x11-drivers/xf86-video-i810
     Available versions:  1.4.1.3 ~1.6.0 ~1.6.1
     Installed:           1.4.1.3
     Homepage:            http://xorg.freedesktop.org/
     Description:         X.Org driver for i810 cards

maybe X doesn't care about the .3 part

As stated earlier - it might be worth to test with 7.1. I've read somewhere that there are only some dependencies - at least much less than upgrading from 6.x. Don't know if 7.0 would work w/ 1.6.x drivers.

I tried an emerge -pv a while ago and it only wanted to upgrade a few packages, so i don't think that it is a big deal. i could simply unmask xf86-video-i810... i could ask people more in the know if this is a no-no.

What makes me wonder is the different behaviuor of mplayer between your machine and my. I assume you are not using KDE what could be one reason. The other one might be flags or gcc. Here is what I have:

you are right that is peculiar. i checked and i have, essentially, the same system (same gcc, same cflags, etc), i really don't know what the cause of it could be. in my case, it could be mplayer over nfs. hmmm, i will think. actually, can you tell me the output of emerge -av mplayer?

[amaroc]

actually, can you tell me the output of emerge -av mplayer?

sure

Code: Select all

# emerge -pv mplayer

These are the packages that would be merged, in order:

Calculating dependencies... done!
[ebuild   R   ] media-video/mplayer-1.0_pre8  USE="X aac alsa arts cdparanoia dvd dvdread encode gif i8x0 ipv6 jpeg mad opengl oss png real sdl theora truetype unicode vorbis win32codecs xmms xv xvid -3dfx -3dnow -3dnowext -aalib -bidi -bindist -bl -cpudetection -custom-cflags -debug -dga -directfb -doc -dts -dv -dvb -esd -fbcon -ggi -gtk -jack -joystick -libcaca -lirc -live -livecd -lzo -matrox -mmx -mmxext -musepack -nas -nvidia -openal -rtc -samba -speex -sse -sse2 -svga -tga -v4l -v4l2 -x264 -xanim -xinerama -xvmc" 0 kB

Total size of downloads: 0 kB

Edit: What BIOS version are you using? Mine is A08 as stated earlier.

[juniper]

i think my bios is A05. should i update? can you do that in linux?

[amaroc]

i think my bios is A05. should i update? can you do that in linux?

You should think about a BIOS update:
- A05 has Intel 915GM Graphics Video BIOS version 3412
- as from A06 it is Intel 915GM Graphics Video BIOS version 1219
- A09 is the newest http://ftp.us.dell.com/bios/I6000A09.EXE
- A09 has only small improvements over A08 (what I have) - so I have not updated yet
- you can find (older) bios versions and some readmes here:ftp://ftp.dell.com/bios
- you can't do it from linux, you need windows or dos, but freedos should do the job

btw: have you seen https://www.dellbatteryprogram.com/Default.aspx? Happily I'm not affected.

[juniper]

You should think about a BIOS update:
- A05 has Intel 915GM Graphics Video BIOS version 3412
- as from A06 it is Intel 915GM Graphics Video BIOS version 1219
- A09 is the newest http://ftp.us.dell.com/bios/I6000A09.EXE
- A09 has only small improvements over A08 (what I have) - so I have not updated yet
- you can find (older) bios versions and some readmes here:ftp://ftp.dell.com/bios
- you can't do it from linux, you need windows or dos, but freedos should do the job

well, luckily for me, i have kept my windows partition around (in fact, just for this occasion). as anti M$ as i am, i figured i wouldn't remove the windows partition until i ACTUALLY needed the space. i still have a ton of free space on my drive, so M$ is safe for now.

This freedos thing looks interesting. i will look into it.

are there any downsides to a bios upgrade? i guess version A08 is working for you.

btw: have you seen https://www.dellbatteryprogram.com/Default.aspx? Happily I'm not affected.

i just read this. UNHAPPILY, i am not affected. may battery capacity is sucking and i wouldn't mind a new battery. i think i will have to shell out for that.

[amaroc]

are there any downsides to a bios upgrade? i guess version A08 is working for you.

Reflashing is a always a bit risky. But there are a lot of improvements over A05, see ftp://ftp.dell.com/bios/I6000A08.TXT Espacially the updated video-bios could be of help.
But - I have not seen an i6000A05.exe in order to reflash if something is worse after the update.

may battery capacity is sucking and i wouldn't mind a new battery. i think i will have to shell out for that.

This could be a reason to flash A09 right now. A09 claims to have a better charging algorithm and has been released almost 1 year ago - so not much additional risk. It will not help your current battery but your new one might last longer.

[juniper]

have you seen libsmbios? apparently, there is a way to flash the bios in linux

http://linux.dell.com/libsmbios/main/index.html

[amaroc]

have you seen libsmbios? apparently, there is a way to flash the bios in linux

http://linux.dell.com/libsmbios/main/index.html

Didn't know that and was curious to play with it. I was able to extract the .hdr from the .exe successfully but there are so much dependencies including a kernel modul that didn't want to play - so I gave up in the end.
Seems to be much easier to boot win or dos ...
Did you made any progress?

[juniper]

@ amaroc:

well, i upgraded to A09 bios. i think i ran mplayer once or twice and it didn't crash. maybe that is the answer.

other than my battery sucking, the STR is great. STD works as well with the kernel and patches you sent me. i had X crash on me once from a wake up, but the system was still alive.

by the way, i am thinking of getting a new battery. dell has some batteries for sale, but ebay has them for about a third of the price. they are labelled new and come with a warranty from some vendors on ebay.

[juniper]

hmmmm. after upgrading my bios to A09, i also upgraded my system. since them my wireless cuts in and out. here is dmesg

Code: Select all

ipw2200: Firmware error detected.  Restarting.
ipw2200: Sysfs 'error' log already exists.
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101
nfs: RPC call returned error 101

because i did the two upgrades near the same time, i can't tell what is causing the problem.

[amaroc]

hmmmm. after upgrading my bios to A09, i also upgraded my system. since them my wireless cuts in and out. here is dmesg

Code: Select all

ipw2200: Firmware error detected.  Restarting.
ipw2200: Sysfs 'error' log already exists.
nfs: RPC call returned error 101
nfs: RPC call returned error 101

because i did the two upgrades near the same time, i can't tell what is causing the problem.

Looks like you are running from one problem into another

OK, firmware error happens from time to time - even with the latest drivers. However, the driver should be able to handle this by restarting/reloading the firmware.

Let me guess - you're running kernel 2.6.16.x and are using builtin ipw2200 v1.0.8 and ieee80211 v1.1.7 - right?
You should also have ipw2200-firmware 3.0.

When using 2.6.16 I would propose to use seperate ipw2200 and ieee80211. I had much better results than with the builtin in the past. However, you should use the stable ipw2200 v1.1.2 rather then the 1.1.3 that somehow make problems with WPA encryption and wpa_supplicant.

Another option is to upgrade your kernel. As you know, I'm running a original 2.6.17-gentoo-r4 kernel that has builtin ipw2200 v1.1.1 and ieee80211 v1.1.7 and have (almost) no problem. The only issue is that wpa_supplicant sometimes needs about 1 minute in order to connect to my WPA-encrypted network. I don't know whether it comes from ipw2200, ieee, wpa_supplicant, ...

[juniper]

alright amaroc. i will give it a try and let you know how it goes.

thanks for all the help.

EDIT: bios update seems to have worked. i haven't had a single video crash since then (i have tried a few times).

another happy note, i just got a new battery. ahhhhhhhhhhhhhhhhhhhhhh, lasts a good 4 hours it seems.

[masterinferno]

I can't make the ipw2200 LED working

I have

Code: Select all

options ipw2200 led=1

in /etc/modules.d/ipw2200
but when I start /etc/init.d/net.eth1 the LED doesn't turn on

Did I forget something ? Do I have to enable LED support in kernel or is it for others devices ? Thanks in advance

[masterinferno]

I must restart /etc/init.d/net.eth1 in order to have the led on...
How can I have it during boot process ?

[m0rd0]

I must restart /etc/init.d/net.eth1 in order to have the led on...
How can I have it during boot process ?

Not sure, I've been wondering about this too. Anyone actually got the LED working on boot?

[amaroc]

I must restart /etc/init.d/net.eth1 in order to have the led on...
How can I have it during boot process ?

Not sure, I've been wondering about this too. Anyone actually got the LED working on boot?

In order to get most of the driver's LED capabilities enabled I suggest to apply the patch from Mark Lord's kernel patch collection for the Dell Inspiron 9300. You still need

Code: Select all

add options ipw2200 led=1

to /etc/modules.conf
The patch seems to be a little bit outdated but still works here with built-in ipw2200 driver from 2.6.17-r8-gentoo.
I can't remember whether it works after boot but it does it for STR where my ipw2200 module gets unloaded and loaded.

Edit: I've recently booted my i6k and would like confirm that the LED-patch works on boot as expected.

[amaroc]

My I6K is now about one year old and it can be said that it is very well supported by gentoo-linux. But as usually - there are still some improvements possible.
One point is that the fan never stops running. Even if it's not really noisy and the fan runs most of the time at lowest speed it would be nice to stop the fan when the laptop is in idle or low activity state.
I've searched a while for such feature and experimented a little bit. Please treat the information underneath as a collection from different sites and forums.

Fan-control
Unfortunately there seems to be no way to control the fan via pwm and lm_sensors - at least I didn't find any. Like other Dell laptops there are only 3 fan-states possible: off, lo- and hi-speed.
There are several tools around for setting one of these three states that seem all derived from a reverse engeneered Dell diagnostic tool
http://www.diefer.de/i8kfan/i8kfan.html

Variant 1)
The most common way is to use the kernel module i8k that offers the lo-level interface to access Dell's SMBIOS functions. You have to enable it via

Code: Select all

CONFIG_I8K=m

and later on to load it via

Code: Select all

modprobe i8k

A user interface is offered via the i8kutils, get it via

Code: Select all

emerge i8kutils

To test the interface you should issue the i8kctl command and get somehow the following

Code: Select all

i6kgt ~ # i8kctl
1.0 (null) 45Y5W1J 44 -1 0 922 0 0 -1

You are now able to set the fan-speed

Code: Select all

i8kctl fan - 0

where '0' disables the fan and 1 or 2 sets the fan to lo- or hi-speed.

Don't forget to

Code: Select all

echo i8k >> /etc/modules.autoload.d/kernel-2.6

Now you need a way to set the fanspeed according to your temperature. You need a script /usr/sbin/i6kfancontrol

Code: Select all

#!/bin/bash
#
INTERVAL=5
MINTEMP=38
MIDTEMP=48
MAXTEMP=58
FANOFF=0
FANMID=1
FANMAX=2

echo 'Starting fan control for i6k ...'

while true
do
	status=`i8kctl`
	temp=`echo $status | cut -d" " -f4`
	fanstate=`echo $status | cut -d" " -f6`	
	fan=$fanstate
	if (( ${temp} < ${MINTEMP} )) ;
	    then fan=$FANOFF
	elif (( ${temp} > ${MAXTEMP} )) ;
	    then fan=$FANMAX
	elif (( ${temp} > ${MIDTEMP} )) ; then
	    if (( ${fanstate} == ${FANOFF} )) ; 
    		then fan=$FANMID
	    fi;
	elif (( ${fanstate} == ${FANMAX} )) ;  
    		then fan=$FANMID
	fi;
	if (( ${fan} != ${fanstate} )) ;
		then i8kctl fan - `echo $fan`
	fi;
	sleep $INTERVAL
done

and something to start at /etc/init.d/i6kfancontrol

Code: Select all

#!/sbin/runscript
# 

PNAME=i6kfancontrol


start() {
	ebegin "Starting i6kfancontrol"
	start-stop-daemon --start --quiet --background \
		--exec /usr/sbin/i6kfancontrol
	eend ${?}
}

stop() {
	ebegin "Stopping i6kfancontrol"
	start-stop-daemon --stop --name ${PNAME}
	eend ${?}
}

Don't forget to

Code: Select all

chmod +x /usr/sbin/i6kfancontrol
rc-update add i6kfancontrol default

Variant 2)
Beside the way with i8k-kernel-module, i8kutils and some bash scripts there is an interesting project here http://dellfand.dinglisch.net/
It uses more or less the same lo-level functions but does not need the kernel-module. Anyway, I like little bash-scripts ...

Note: Even if you read somewhere that aborting the script will leave you with an unregulated fan - my i6k doesn't behave so. The BIOS is still controlling the fan if you reach critical temperatures. On the other hand, you shouldn't overrule the tresholds from the BIOS - otherwise your notebook might get overheated and permanently damaged - don't say I didn't warn you.

This leads us to other issues: temperature, power consumption and battery life-time.

Power management
The Pentium-M cpu offers the so called speed stepping. In short - it is an adaptation of the cpu-clock to the current system-load and an adaptation of the cpu-voltage to the cpu-clock in parallel. The linux-kernel offers this functionality by

Code: Select all

CONFIG_X86_SPEEDSTEP_CENTRINO=y
CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI=y
CONFIG_X86_SPEEDSTEP_CENTRINO_TABLE=y
...
CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
CONFIG_CPU_FREQ_GOV_POWERSAVE=y
CONFIG_CPU_FREQ_GOV_USERSPACE=y
CONFIG_CPU_FREQ_GOV_ONDEMAND=y
CONFIG_CPU_FREQ_GOV_CONSERVATIVE=y

There are a lot of guides and howto's around that describe how to setup frequency-scaling and the different governors you want to look for.
The ondemand governor seems to offer the best trade-off between performance and battery-lifetime though.
However, the frequency-scaling comes along with the default cpu-voltage settings from the BIOS that are often much too conservative. So it's worth of playing a little bit with it.

Undervolting

There are differnt approaches to adopt the speed-step voltages. One way is to patch the kernel to provide those voltage tables for the kernel's speed-stepping governors. The other and I think more flexible way is the user-space. One really great tool is cpupw.

Download cpupw from:
http://www.tuxamito.com.es/cpupw/index.php
and see also
http://forums.gentoo.org/viewtopic-t-44 ... fc45b080f4
http://forums.gentoo.org/viewtopic-t-44 ... fc45b080f4

then untar and install

Code: Select all

cd src
make
su
make install

Make sure your kernel has msr support.

Code: Select all

CONFIG_X86_MSR=m

and the needed governors are available, e.g.

Code: Select all

CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE=y
CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
CONFIG_CPU_FREQ_GOV_POWERSAVE=y
CONFIG_CPU_FREQ_GOV_USERSPACE=y
CONFIG_CPU_FREQ_GOV_ONDEMAND=y
CONFIG_CPU_FREQ_GOV_CONSERVATIVE=y

then

Code: Select all

modprobe msr
echo msr >> /etc/modules.autoload.d/kernel-2.6

and a

Code: Select all

cpupw --status

should show something usefull

Now you need a VID configuration for your CPU. Here's what the author wrote:
http://forums.gentoo.org/viewtopic-p-31 ... ml#3198400

That's the default of my I6K and Pentium M 1.6 GHz:

Code: Select all

1
1
592
6
1600000 40 -1 -1
1400000 36 -1 -1
1200000 32 -1 -1
1000000 27 -1 -1
800000 23 -1 -1
600000 18 -1 -1
1

that I've changed to

Code: Select all

1
1
592
6
1600000 27 -1 -1
1400000 23 -1 -1
1200000 19 -1 -1
1000000 14 -1 -1
800000 10 -1 -1
600000 5 -1 -1
1

what seems to be stable because I'm running it for some weeks now. I've stored it as /etc/cpupw/config.
You can now test and fine tune a little bit by

Code: Select all

/usr/sbin/cpupw --config /etc/cpupw/config --smart -g userspace -t 400 -u 80 -D 45 -f

I've adopted the -t xxx parameter a little bit. It's the time that cpupw sleeps between the internal measures. If you run cpupw in foreground mode it shows any change regarding the frequency. In order to get the longest battery life-time it's a good idea to make it as little as possible for your preferred idle state (e.g. kde desktop) without any frequency change. Thus you won't get any switching on idle and fastest possible performance adoptation.
You need again something to start at /etc/init.d/cpupw

Code: Select all

#!/sbin/runscript
# 

CONFIG=/etc/cpupw/config
PNAME=cpupw

checkconfig() {
	if [ ! -f ${CONFIG} ]; then
		eerror "Configuration file ${CONFIG} not found"
		return 1
	fi
}

start() {
	checkconfig || return 1

	ebegin "Starting cpupw"
	start-stop-daemon --start --quiet --background \
		--exec /usr/sbin/cpupw -- --config ${CONFIG} --smart -g userspace -t 400 -u 80 -D 45 -d
	eend ${?}
}

stop() {
	ebegin "Stopping cpupw"
	start-stop-daemon --stop --name ${PNAME}
	eend ${?}
}

and don't forget to

Code: Select all

rc-update add cpupw default

Result
Using the fan-control only the i6k stays quite almost 80% of the time in idle-mode. However, even web-browsing or email-read/write forces the fan to spin from time to time. The average power consumption is 17W (cat /proc/acpi/battery/BAT0/state). The temperature drifts between the two tresholds of 38 and 48°C.
Under heavy load like

Code: Select all

emerge openoffice

or

Code: Select all

/opt/gimps/mprime -t -w/tmp

the fan runs at full speed all the time and the power consumption is about 33W. The temperature stays at 62°C.
Summary default:
Idle -> 17W, 38...48°C, 80% fan off
Load -> 33W, 62°C, 100% fan at full speed

The same scenarious with cpupw:
Idle -> 15W, 42°C, 100% fan off
Load -> 25W, 52°C, 100% fan at lo-speed

Even with web-browsing then fan stays off almost all the time and never reaches full-speed under heavy load. Also, I could not see a measurable perfomance impact due to the userspace governor from cpupw. See as an example gdb-compile 1st with cpupw and 2nd with ondemand governor:

Code: Select all

i6kgt ~ # genlop --date 1 hour ago -t gdb
 * sys-devel/gdb

     Mon Oct 23 23:36:10 2006 >>> sys-devel/gdb-6.4
       merge time: 4 minutes and 33 seconds.

     Mon Oct 23 23:42:02 2006 >>> sys-devel/gdb-6.4
       merge time: 4 minutes and 33 seconds.

For performance comparison with my Pentium-M @1.6GHz, 400MHz FSB

Code: Select all

i6kgt ~ # genlop -t openoffice
     Sun Oct 22 17:31:44 2006 >>> app-office/openoffice-2.0.3
       merge time: 4 hours, 44 minutes and 37 seconds.

Harddisk spin-down
If you are at this state you might find the harddisk noisy. However, this depends heavily on your harddisk and your personal feeling - or age if you want:-) My Samsung HM120JC isn't very noisy and I can't really see much different on overall power consumption if I spin it down via laptop-mode from http://www.samwel.tk/laptop_mode/index.html

Code: Select all

/etc/init.d/laptop_mode start

The difference seems to be about 0.5W - not considering the extra amount of power for the spin-up from time to time. I personally have decided not to use laptop-mode.

If you have reached this line you seem to have a lot of time and it might be worth of playing with fan-control and under-voltage - so have fun

EDIT 20070107:
- After running this configuration over about 3 months it can be confirmed as really stable. The fan is almost all the time off.
- removed "errno.h" include for dellfand - comes with latest sources

[masterinferno]

What about making an article on gentoo-wiki ?

[entrari]

hi,
I have an INspiron 6000d with 128 ati x300 video card
i can't seem to boot it up

it works when its on safe mode so i tried uninstalling the video card
without the video card driver it works in normal mode.
at first i'd assumed the video card was broken but when i used an external monitor it booted ok.

ive tried all the latest drivers from dell and ati to no avail.

[pblinux2]

Does anyone have a good xorg.conf for modular X on an Inspiron 6000 with the M300 and 1280x800 panel? I just spend 6 hours going back and forth over this thread, and I can't tell what is still relevant and what is outdated. I'm beating my head against the wall on this.

[seppelrockt]

I must restart /etc/init.d/net.eth1 in order to have the led on...
How can I have it during boot process ?

Not sure, I've been wondering about this too. Anyone actually got the LED working on boot?

In order to get most of the driver's LED capabilities enabled I suggest to apply the patch from Mark Lord's kernel patch collection for the Dell Inspiron 9300. You still need

Code: Select all

add options ipw2200 led=1

to /etc/modules.conf
The patch seems to be a little bit outdated but still works here with built-in ipw2200 driver from 2.6.17-r8-gentoo.
I can't remember whether it works after boot but it does it for STR where my ipw2200 module gets unloaded and loaded.

Edit: I've recently booted my i6k and would like confirm that the LED-patch works on boot as expected.

Does this mean the LED is on whenever WLAN is activated via Fn+F2, even when you are not connected to an AP? This doesn't seem to work here with ipw2200-1.1.3 (regardless of the patch you recommented). I would love to get this working cause I often forget to disable WLAN when I am at home (there are no APs available) and this costs at least some battery time.

[amaroc]

Does this mean the LED is on whenever WLAN is activated via Fn+F2, even when you are not connected to an AP?

Not exactly - there are three states
- OFF: disabled via Fn+F2
- Blinking: Negotiation with AP
- ON: Connected

Here is what var/log/messages reports after a cycle Fn+F2 / Fn+F2

Code: Select all

Nov 23 20:45:09 i6kgt atkbd.c: Unknown key pressed (translated set 2, code 0x88 on isa0060/serio0).
Nov 23 20:45:09 i6kgt atkbd.c: Use 'setkeycodes e008 <keycode>' to make it known.
Nov 23 20:45:10 i6kgt wpa_cli: interface eth1 DISCONNECTED
Nov 23 20:46:27 i6kgt atkbd.c: Unknown key pressed (translated set 2, code 0x88 on isa0060/serio0).
Nov 23 20:46:27 i6kgt atkbd.c: Use 'setkeycodes e008 <keycode>' to make it known.
Nov 23 20:46:29 i6kgt ipw2200: Failed to send CARD_DISABLE: Command timed out.
Nov 23 20:46:29 i6kgt wpa_cli: interface eth1 CONNECTED
Nov 23 20:46:36 i6kgt dhclient: DHCPREQUEST on eth1 to 255.255.255.255 port 67
Nov 23 20:46:36 i6kgt dhclient: DHCPACK from 192.168.65.1
Nov 23 20:46:36 i6kgt dhclient: bound to 192.168.65.3 -- renewal in 3259978 seconds.

I don't know whether the LED would finish to blink sometime if no AP can be seen.

This doesn't seem to work here with ipw2200-1.1.3 (regardless of the patch you recommented). I would love to get this working cause I often forget to disable WLAN when I am at home (there are no APs available) and this costs at least some battery time.

For me it works with kernel built-in ipw2200 and the patch mentioned on both
- linux-2.6.17-gentoo-r8 (ipw2200 v1.1.1)
- linux-2.6.18-gentoo-r2 (ipw2200 v1.1.2)
Need to mention that 2.6.18 needed the wpa_supplicant parameter
ap_scan=2
in order to connect to my WPA-PSK AP (SE515DSL).