ACPI + Laptop Power Management

ACPI + Laptop Power Management

[Ow Mun Heng]

Hi,

	This is a question (not a solution) to determine how to gain maximum
batt life on a laptop.

More details on http://bellet.info/~bellet/laptop/t40.html

Currently, following the recommendations in the link above, I have set
'noatime' for my HDs as well as placing commit=600.

so in /etc/fstab

/dev/hda	/	etx3	defaults,noatime,commit=600 1 1

The commit=600 is supposed to stop Kjournal from committing to disk every 5
seconds and to give a chance to the HD to spin down (not happening!) (can
check via hdparm -C /dev/hda)

Doing a hdparm -y /dev/hda (to place the disk into standby) will only work
for like 10-30 seconds before it spins up again. I have no idea what is
causing it to spin up again.. (what I/O is going on??)

There is the option of changing bdflush as well, changing this (5th & 6th)
to 6000 & 12000 does not yield anything better

Does anyone has a solution or something like a solution which I can try
out??

Anything would be good.. I want to extend my battery life.. Dell D600 on a
4320Mah 6cell Batt with 14.1 SXGA screen

Cheers,                                                 .^.
Mun Heng, Ow                                            /V\
H/M Engineering                                       /(   )\
Western Digital M'sia                                  ^^-^^
DID : 03-7870 5168                          The Linux Advocate

        


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Ducrot Bruno]

On Fri, Nov 21, 2003 at 12:01:03PM +0800, Ow Mun Heng wrote:
> Hi,
> 
> 	This is a question (not a solution) to determine how to gain maximum
> batt life on a laptop.
> 
> More details on http://bellet.info/~bellet/laptop/t40.html
> 
> Currently, following the recommendations in the link above, I have set
> 'noatime' for my HDs as well as placing commit=600.
> 
> so in /etc/fstab
> 
> /dev/hda	/	etx3	defaults,noatime,commit=600 1 1
> 
> The commit=600 is supposed to stop Kjournal from committing to disk every 5
> seconds and to give a chance to the HD to spin down (not happening!) (can
> check via hdparm -C /dev/hda)
> 
> Doing a hdparm -y /dev/hda (to place the disk into standby) will only work
> for like 10-30 seconds before it spins up again. I have no idea what is
> causing it to spin up again.. (what I/O is going on??)
> 
> There is the option of changing bdflush as well, changing this (5th & 6th)
> to 6000 & 12000 does not yield anything better
> 

I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
You should really look at.  Look at linux/Documentation/laptop-mode.txt
for more details.


-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

on 11/21/03 11:12, Ducrot Bruno wrote:
| I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
| You should really look at.  Look at linux/Documentation/laptop-mode.txt
| for more details.
sorry, I read the doc but I couldn't found all the info I need... :-(

~From the doc above...
| The main knob is /proc/sys/vm/laptop_mode. Setting that to 1 switches the
| vm (and block layer) to laptop mode. Leaving it to 0 makes the kernel work
| like before. When in laptop mode, you also want to extend the intervals
| desribed above. See the laptop-mode.sh script for how to do that.
1) As I understood, to enable 'laptop_mode' I just need to do an
	echo 1 > /proc/sys/vm/laptop_mode
right? I mean, it's already present in the kernel, I don't need to
enable any new features in kernel config or somewhere else, right?

2) Where I can find 'laptop-mode.sh'??? It's not present in kernel
source dir...

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/ve+yVAp7Xm10JmkRAhbmAJ9BwYbGTnQAGKaT66Ixs4PqAi1k+wCaAv4z
YbadY+gQZi/2b/Eiex+APZY=
=TvIN
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

[-- Attachment #1: Type: text/plain, Size: 2755 bytes --]

On Fri, Nov 21, 2003 at 11:12:32AM +0100, Ducrot Bruno wrote:
> On Fri, Nov 21, 2003 at 12:01:03PM +0800, Ow Mun Heng wrote:
> > Hi,
> > 
> > 	This is a question (not a solution) to determine how to gain maximum
> > batt life on a laptop.
> > 
> > More details on http://bellet.info/~bellet/laptop/t40.html
> > 
> > Currently, following the recommendations in the link above, I have set
> > 'noatime' for my HDs as well as placing commit=600.
> > 
> > so in /etc/fstab
> > 
> > /dev/hda	/	etx3	defaults,noatime,commit=600 1 1
> > 
> > The commit=600 is supposed to stop Kjournal from committing to disk every 5
> > seconds and to give a chance to the HD to spin down (not happening!) (can
> > check via hdparm -C /dev/hda)
> > 
> > Doing a hdparm -y /dev/hda (to place the disk into standby) will only work
> > for like 10-30 seconds before it spins up again. I have no idea what is
> > causing it to spin up again.. (what I/O is going on??)
> > 
> > There is the option of changing bdflush as well, changing this (5th & 6th)
> > to 6000 & 12000 does not yield anything better
> > 
> 
> I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
> You should really look at.  Look at linux/Documentation/laptop-mode.txt
> for more details.
>

like said here, laptopmode is what you want since it holds back dirty
buffer flushes which is what you are experiencing. It will fire up the
disk on reads though. I attached a script that you can use to enter
laptopmode (was actually distributed with laptopmode as far as I could
tell when googling for the original patch). I call it in responce to
ac events from acpid to get it running when the power is disconnected.
Will have to make sure that you get those event though since I had to
twik my dsdt for that.
You can also look at cpufreq of acpi cpu throttling to throttle down you
cpu when on battery to conserve power (assuming you cpu supports that).
I patched reiserfs to change the journal commit time according to the entry in
bdflush so that I can get the regular 30 sec journal flush time when on
ac and 10 minutes when of battery.
I could try making such a patch for ext3 if you want.

> 
> -- 
> Ducrot Bruno
> 
> --  Which is worse:  ignorance or apathy?
> --  Don't know.  Don't care.
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive?  Does it
> help you create better code?  SHARE THE LOVE, and help us help
> YOU!  Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 

[-- Attachment #2: laptopmode_script --]
[-- Type: text/plain, Size: 1260 bytes --]

#!/bin/sh
# start or stop laptop mode. best run by a power management daemon when ac gets
# connected/disconnected from a laptop.

# FIXME : assumes HZ * <age time in seconds> (HZ = 100). should be put into a
# sysconfig file

# The dirty buffer age and update time when running on laptop mode
MAX_AGE=600

# The kernel's default dirty buffer age and update time
DEF_AGE=30
DEF_UPDATE=5

if [ ! -w /proc/sys/vm/laptop_mode ]; then
    echo "Kernel is not patched with laptop_mode patch"
    exit 1
fi

case "$1" in
start)
    AGE=$((100*$MAX_AGE))

    echo -n "Starting laptop mode"
    echo "1" > /proc/sys/vm/laptop_mode
    echo "30 500 0 0 $AGE $AGE 60 20 0" > /proc/sys/vm/bdflush

    # if no AAM support, manually set spin down
    if [ ! -d /proc/aam/hda ]; then
	hdparm -S6 /dev/hda
    else
	echo 1 > /proc/aam/hda/enabled
    fi

    echo ".";;

stop)
    U_AGE=$((100*$DEF_UPDATE))
    B_AGE=$((100*$DEF_AGE))

    echo -n "Stopping laptop mode"
    echo "0" > /proc/sys/vm/laptop_mode
    echo "30 500 0 0 $U_AGE $B_AGE 60 20 0" > /proc/sys/vm/bdflush

    # Disable spin down
    if [ ! -d /proc/aam/hda ]; then
	hdparm -S0 /dev/hda
    else
	echo 0 > /proc/aam/hda/enabled
    fi

    echo ".";;

*)
    echo "$0 {start|stop }";;

esac
exit 0

Re: ACPI + Laptop Power Management

[Ducrot Bruno]

[-- Attachment #1: Type: text/plain, Size: 1549 bytes --]

On Fri, Nov 21, 2003 at 11:57:55AM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello,
> 
> on 11/21/03 11:12, Ducrot Bruno wrote:
> | I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
> | You should really look at.  Look at linux/Documentation/laptop-mode.txt
> | for more details.
> sorry, I read the doc but I couldn't found all the info I need... :-(
> 
> ~From the doc above...
> | The main knob is /proc/sys/vm/laptop_mode. Setting that to 1 switches the
> | vm (and block layer) to laptop mode. Leaving it to 0 makes the kernel work
> | like before. When in laptop mode, you also want to extend the intervals
> | desribed above. See the laptop-mode.sh script for how to do that.
> 1) As I understood, to enable 'laptop_mode' I just need to do an
> 	echo 1 > /proc/sys/vm/laptop_mode
> right? I mean, it's already present in the kernel, I don't need to
> enable any new features in kernel config or somewhere else, right?

No need to add a config.  But just echoing to the laptop sysctl is not
engough though.  You have to tweak other VM stuff for it to be fully
functional.

> 2) Where I can find 'laptop-mode.sh'??? It's not present in kernel
> source dir...

Ah, hem, that seems to be missing...  Here is the orignal one posted,
or from a link (I don't remember) to LKML.
I use it with another script to handle AC change via acpid.
It is also called with 'stop' on shutdwon, or reboot.


Cheers,

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.

[-- Attachment #2: laptop-mode --]
[-- Type: text/plain, Size: 856 bytes --]

#!/bin/sh
#
# start of stop laptop mode, best run by a power management daemon when
# ac gets connected/disconnected from a laptop
#
# FIXME: assumes HZ == 100

# age time, in seconds. should be put into a sysconfig file
MAX_AGE=600

# kernel default dirty buffer age
DEF_AGE=30
DEF_UPDATE=5

if [ ! -w /proc/sys/vm/laptop_mode ]; then
	echo "Kernel is not patched with laptop_mode patch"
	exit 1
fi

case "$1" in
	start)
		AGE=$((100*$MAX_AGE))
		echo -n "Starting laptop mode"
		echo "1" > /proc/sys/vm/laptop_mode
		echo "30 500 0 0 $AGE $AGE 60 20 0" > /proc/sys/vm/bdflush
		echo "."
		;;
	stop)
		U_AGE=$((100*$DEF_UPDATE))
		B_AGE=$((100*$DEF_AGE))
		echo -n "Stopping laptop mode"
		echo "0" > /proc/sys/vm/laptop_mode
		echo "30 500 0 0 $U_AGE $B_AGE 60 20 0" > /proc/sys/vm/bdflush
		echo "."
		;;
	*)
		echo "$0 {start|stop}"
		;;

esac

exit 0

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Micha,

on 11/21/03 12:13, Micha Feigin wrote:
| like said here, laptopmode is what you want since it holds back dirty
| buffer flushes which is what you are experiencing. It will fire up the
| disk on reads though. I attached a script that you can use to enter
| laptopmode (was actually distributed with laptopmode as far as I could
| tell when googling for the original patch). I call it in responce to
| ac events from acpid to get it running when the power is disconnected.
| Will have to make sure that you get those event though since I had to
| twik my dsdt for that.
thx for the script, I was searching for but I couldn't find :-(
On the other hand, what do I search for in my DSDT? I mean, in which way
can I know if the 'laptop_mode' is supported on my DSDT?

| I could try making such a patch for ext3 if you want.
Yes, please :-)

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/vg4AVAp7Xm10JmkRAmsWAJ928z25MNHuka49V08QWHa0NryAxACdFX8I
iEg+mpKXPRz2VGqOOmu8y3g=
=liFy
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Ducrot Bruno]

On Fri, Nov 21, 2003 at 02:07:14PM +0100, Luca Capello wrote:
> Hello Micha,
> 
> on 11/21/03 12:13, Micha Feigin wrote:
> | like said here, laptopmode is what you want since it holds back dirty
> | buffer flushes which is what you are experiencing. It will fire up the
> | disk on reads though. I attached a script that you can use to enter
> | laptopmode (was actually distributed with laptopmode as far as I could
> | tell when googling for the original patch). I call it in responce to
> | ac events from acpid to get it running when the power is disconnected.
> | Will have to make sure that you get those event though since I had to
> | twik my dsdt for that.
> thx for the script, I was searching for but I couldn't find :-(
> On the other hand, what do I search for in my DSDT? I mean, in which way
> can I know if the 'laptop_mode' is supported on my DSDT?

laptop-mode and ACPI are complete different things...
This work like a charm in my iBook for instance.

> 
> | I could try making such a patch for ext3 if you want.
> Yes, please :-)
> 

If you active the laptop-mode, then commit interval for ext3 will change
to the sysctl buffer_age, so if that one is changed to be 600 (like in
the script), then commit intervall will auto-magically be 10 min.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

On Fri, Nov 21, 2003 at 02:07:14PM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello Micha,
> 
> on 11/21/03 12:13, Micha Feigin wrote:
> | like said here, laptopmode is what you want since it holds back dirty
> | buffer flushes which is what you are experiencing. It will fire up the
> | disk on reads though. I attached a script that you can use to enter
> | laptopmode (was actually distributed with laptopmode as far as I could
> | tell when googling for the original patch). I call it in responce to
> | ac events from acpid to get it running when the power is disconnected.
> | Will have to make sure that you get those event though since I had to
> | twik my dsdt for that.
> thx for the script, I was searching for but I couldn't find :-(
> On the other hand, what do I search for in my DSDT? I mean, in which way
> can I know if the 'laptop_mode' is supported on my DSDT?

laptopmode has nothing to do with you dsdt. The dsdt is the acpi code in
the bios. Its relevent to look into it if you are using acpi and having
trouble since a lot of the bioses have a buggy dsdt table (acpi code
that is).
If you are using acpi look into acpid to be able to act upon acpi
events.
I this case, with acpid not running do (assuming you do have acpi
support, otherwise you won't see this file)
cat /proc/acpi/event
and pull out your power cord and then reinsert it. If you are getting
ACAD or BAT or somesuch thing messages then everything is ok (I don't
know how to do this with apm) and you can use acpid to run the
laptopmode script whenever your power is connected/disconnected.


> 
> | I could try making such a patch for ext3 if you want.
> Yes, please :-)

Will look into it.

> 
> Thx, bye,
> Gismo / Luca
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
> Comment: Using GnuPG with Debian - http://enigmail.mozdev.org
> 
> iD8DBQE/vg4AVAp7Xm10JmkRAmsWAJ928z25MNHuka49V08QWHa0NryAxACdFX8I
> iEg+mpKXPRz2VGqOOmu8y3g=
> =liFy
> -----END PGP SIGNATURE-----
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive?  Does it
> help you create better code?  SHARE THE LOVE, and help us help
> YOU!  Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Micha,

on 11/21/03 14:53, Micha Feigin wrote:
| If you are using acpi look into acpid to be able to act upon acpi
| events.
| I this case, with acpid not running do (assuming you do have acpi
| support, otherwise you won't see this file)
| cat /proc/acpi/event
| and pull out your power cord and then reinsert it. If you are getting
| ACAD or BAT or somesuch thing messages then everything is ok (I don't
| know how to do this with apm) and you can use acpid to run the
| laptopmode script whenever your power is connected/disconnected.
yes, I can, I mean, I had already checked for this event (AC-adapter)
and it works :-) BTW, for those who don't know, I'm using an ASUS M3N
(Centrino) with the latest BIOS available (0205A released 20031117).
I've just a little problem and I'm going to inform this list ASAP (I
need some tests before).

| Will look into it.
Thank you very much!

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/viTaVAp7Xm10JmkRAuHGAJ910nA2C57ZEyGlZ2ymRZQpTBwFjQCfXlIy
xCTsJEBr1k1Pqb+LvwC11rw=
=+fWR
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

on 11/21/03 14:03, Ducrot Bruno wrote:
| No need to add a config.  But just echoing to the laptop sysctl is not
| engough though.  You have to tweak other VM stuff for it to be fully
| functional.
The other VM stuffs are included in the 'laptop-mode' scritp, right? I
mean, the script makes all the work for me when I start it or do I need
something other?

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/v6uJVAp7Xm10JmkRAkRTAJ4xxrZ8K3JZdau3TFAmVnBP5vVqXQCgha7K
+eX76aKEj1F4uUo4g/T3o+Q=
=eGiy
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

On Sat, Nov 22, 2003 at 07:31:38PM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello,
> 
> on 11/21/03 14:03, Ducrot Bruno wrote:
> | No need to add a config.  But just echoing to the laptop sysctl is not
> | engough though.  You have to tweak other VM stuff for it to be fully
> | functional.
> The other VM stuffs are included in the 'laptop-mode' scritp, right? I
> mean, the script makes all the work for me when I start it or do I need
> something other?
> 

The script does everything related to laptopmode as long as the kernel
support exists.
You may also want to look into cpufreq or cpu throttling through acpi
which the script doesn't do (requires both kernel and bios/cpu support).

> Thx, bye,
> Gismo / Luca
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
> Comment: Using GnuPG with Debian - http://enigmail.mozdev.org
> 
> iD8DBQE/v6uJVAp7Xm10JmkRAkRTAJ4xxrZ8K3JZdau3TFAmVnBP5vVqXQCgha7K
> +eX76aKEj1F4uUo4g/T3o+Q=
> =eGiy
> -----END PGP SIGNATURE-----
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive?  Does it
> help you create better code?  SHARE THE LOVE, and help us help
> YOU!  Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

on 11/21/03 11:12, Ducrot Bruno wrote:
| I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
| You should really look at.  Look at linux/Documentation/laptop-mode.txt
| for more details.
I saw that in kernel 2.6.0-test9 'laptop_mode' is not present: is there
a patch or something similar to enable it? Or is it merged in something
other?

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/wQatVAp7Xm10JmkRAglIAJ44kdtOlYwWEHwzmSXAP0624KHFfgCghtKD
cKYcNBy/5sLn58KV7GXgRPE=
=SrPG
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Ducrot Bruno]

On Sun, Nov 23, 2003 at 08:12:46PM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello,
> 
> on 11/21/03 11:12, Ducrot Bruno wrote:
> | I'm using laptop-mode which is now in linux-2.4 since 2.4.23-pre7.
> | You should really look at.  Look at linux/Documentation/laptop-mode.txt
> | for more details.
> I saw that in kernel 2.6.0-test9 'laptop_mode' is not present: is there
> a patch or something similar to enable it? Or is it merged in something
> other?
> 

AFAIK, laptop-mode is only in 2.4, and I am not aware of a port
to 2.6.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Luca Capello]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

sorry boring you guys with such a lot of questions...

on 11/24/03 13:44, Ducrot Bruno wrote:
| AFAIK, laptop-mode is only in 2.4, and I am not aware of a port
| to 2.6.
Is there something similar on kernel 2.6? Because ATM I can have
'cpufreq' support on kernel 2.6, but AFAIK not yet on 2.4.23-rc3, so
there're some power management (PM) options available for one specific
kernel and not the other and viceversa...

Regards,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQE/wg0uVAp7Xm10JmkRAn3nAJ0bRH4EZ/WIMrXzDaIKtSa0GKSRoACdE+tb
696MGbvxsSfL3D4/SdzbPO4=
=AZ5E
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

On Fri, Nov 21, 2003 at 03:44:44PM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello Micha,
> 
> on 11/21/03 14:53, Micha Feigin wrote:
> | If you are using acpi look into acpid to be able to act upon acpi
> | events.
> | I this case, with acpid not running do (assuming you do have acpi
> | support, otherwise you won't see this file)
> | cat /proc/acpi/event
> | and pull out your power cord and then reinsert it. If you are getting
> | ACAD or BAT or somesuch thing messages then everything is ok (I don't
> | know how to do this with apm) and you can use acpid to run the
> | laptopmode script whenever your power is connected/disconnected.
> yes, I can, I mean, I had already checked for this event (AC-adapter)
> and it works :-) BTW, for those who don't know, I'm using an ASUS M3N
> (Centrino) with the latest BIOS available (0205A released 20031117).
> I've just a little problem and I'm going to inform this list ASAP (I
> need some tests before).
> 
> | Will look into it.
> Thank you very much!
> 

Someone already answeared that, but I looked into it and ext3 already
takes into account the entries in bdflush to postpone flushing when in
laptop mode (I haven't checked if it checks for laptop mode to do that
but I doubt).
I read in the descripition that ext3 was already handled, but the code
was in the jbd directory and not the ext3 directory, and since I haven't
bothered with ext3 before (I am currently using reiserfs) I thought it
was related to something else.
If anyone wants a similar patch against reiserfs, I got one (if someone
can point me in the right direction to post it so that it can be
reviewed and put into the kernel it will also be nice - is lkml the
place?).

> Thx, bye,
> Gismo / Luca
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
> Comment: Using GnuPG with Debian - http://enigmail.mozdev.org
> 
> iD8DBQE/viTaVAp7Xm10JmkRAuHGAJ910nA2C57ZEyGlZ2ymRZQpTBwFjQCfXlIy
> xCTsJEBr1k1Pqb+LvwC11rw=
> =+fWR
> -----END PGP SIGNATURE-----
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive?  Does it
> help you create better code?  SHARE THE LOVE, and help us help
> YOU!  Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

On Mon, Nov 24, 2003 at 02:52:47PM +0100, Luca Capello wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello,
> 
> sorry boring you guys with such a lot of questions...
> 
> on 11/24/03 13:44, Ducrot Bruno wrote:
> | AFAIK, laptop-mode is only in 2.4, and I am not aware of a port
> | to 2.6.
> Is there something similar on kernel 2.6? Because ATM I can have
> 'cpufreq' support on kernel 2.6, but AFAIK not yet on 2.4.23-rc3, so
> there're some power management (PM) options available for one specific
> kernel and not the other and viceversa...
> 

There is a cpufreq patch for 2.4. Last time I checked it was aimed at
2.4.22 and required some reject fixing for 2.4.23.
Don't know how stable it is with the 2.4.23 kernels (I had some troubles
in conjuction with swsusp and haven't located the exact source yet).
If you are interested in this I can dig up where to download it and a
patch I wrote to fix the rejects (it was for pre8 but probably won't
need too much work).

> Regards,
> Gismo / Luca
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
> Comment: Using GnuPG with Debian - http://enigmail.mozdev.org
> 
> iD8DBQE/wg0uVAp7Xm10JmkRAn3nAJ0bRH4EZ/WIMrXzDaIKtSa0GKSRoACdE+tb
> 696MGbvxsSfL3D4/SdzbPO4=
> =AZ5E
> -----END PGP SIGNATURE-----
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SF.net Giveback Program.
> Does SourceForge.net help you be more productive?  Does it
> help you create better code?  SHARE THE LOVE, and help us help
> YOU!  Click Here: http://sourceforge.net/donate/
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

RE: ACPI + Laptop Power Management

[Ow Mun Heng]

> -----Original Message-----
> From: Micha Feigin [mailto:michf-+lLcF8/aw9x6auLlOhE+pQ@public.gmane.org]
> Sent: Friday, November 21, 2003 7:13 PM
> To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> Subject: Re: [ACPI] ACPI + Laptop Power Management
> 
> 
> On Fri, Nov 21, 2003 at 11:12:32AM +0100, Ducrot Bruno wrote:
> like said here, laptopmode is what you want since it holds back dirty
> buffer flushes which is what you are experiencing. It will fire up the
> disk on reads though. I attached a script that you can use to enter
> laptopmode (was actually distributed with laptopmode as far as I could
> tell when googling for the original patch). I call it in responce to
> ac events from acpid to get it running when the power is disconnected.
> Will have to make sure that you get those event though since I had to
> twik my dsdt for that.
> You can also look at cpufreq of acpi cpu throttling to 
> throttle down you
> cpu when on battery to conserve power (assuming you cpu 
> supports that).

I took a look at laptop mode. However, as the author said "it adds a
non-significant" amount of battery time.

The thing here is, even though it's flushing the buffers when there are
others to flush, it makes not difference if the drive is still spinning and
not in standby-mode! I can't figure out what is making the HD _not_ spin
down.

BTW, Laptop_mode is merged into the 2.4.23-rc5 kernel (based on the
changelogs)

With regard to ACPId events, can you tell me how yo get it to work? I'm
having problems getting it running. It gets lodaded up during init, and it's
running as root, but after logging in (GDM desktop), the events won't fire
until I kill the acpid process and re-run it again. (once I logged in)

I can't figure out how to get it to work. I read in one page where the
auther said to put xhost +local somewhere at the end of /etc/X11/xinitrc.
That didn't work either.

> I patched reiserfs to change the journal commit time 
> according to the entry in
> bdflush so that I can get the regular 30 sec journal flush 
> time when on
> ac and 10 minutes when of battery.
> I could try making such a patch for ext3 if you want.

You mean that has to be patched at the Kernel Code Level?? Ah.. Dang!


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

RE: ACPI + Laptop Power Management

[Diego SANTA CRUZ]

On Fri, 2003-11-28 at 04:21, Ow Mun Heng wrote:

> The thing here is, even though it's flushing the buffers when there are
> others to flush, it makes not difference if the drive is still spinning and
> not in standby-mode! I can't figure out what is making the HD _not_ spin
> down.
> 

Just a thing to check: syslogd will (by default) sync the log file after
each write. If you have a daemon that write syslog messages often then
syslogd will keep your HD active all the time, despite any laptop-mode.

Two, possibly complementary, solutions

1) Find the offending daemon(s) and modify their settings or disable
them if not necessary.

2) Prepend the log file name with a - on the /etc/syslog.conf file. This
prevents the sync on every write default.

If this does not solve your problem you can check (in dmesg) what is
writing to disk by doing

echo -n 1 > /proc/sys/vm/block_dump

But BE CAREFUL!. You have to disable klogd before doing this or you will
likely end up in an infinite loop.

Hope this helps.
-- 
-------------------------------------------------------
Diego Santa Cruz - PhD
Publications available at http://ltswww.epfl.ch/~dsanta
Signal Processing Institute (LTS1 / ITS / STI)
Swiss Federal Institute of Technology (EPFL)
EPFL - STI - ITS, CH-1015 Lausanne, Switzerland
E-mail:     Diego.SantaCruz-p8DiymsW2f8@public.gmane.org
Phone:      +41 - 21 - 693 26 57
Fax:        +41 - 21 - 693 76 00
-------------------------------------------------------



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Ducrot Bruno]

On Fri, Nov 28, 2003 at 11:21:19AM +0800, Ow Mun Heng wrote:
> 
> 
> > -----Original Message-----
> > From: Micha Feigin [mailto:michf-+lLcF8/aw9x6auLlOhE+pQ@public.gmane.org]
> > Sent: Friday, November 21, 2003 7:13 PM
> > To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> > Subject: Re: [ACPI] ACPI + Laptop Power Management
> > 
> > 
> > On Fri, Nov 21, 2003 at 11:12:32AM +0100, Ducrot Bruno wrote:
                                              ^^^^^^^^^^^^

I *never* wrote the following (but there are good stuff
inside).

> > like said here, laptopmode is what you want since it holds back dirty
> > buffer flushes which is what you are experiencing. It will fire up the
> > disk on reads though. I attached a script that you can use to enter
> > laptopmode (was actually distributed with laptopmode as far as I could
> > tell when googling for the original patch). I call it in responce to
> > ac events from acpid to get it running when the power is disconnected.
> > Will have to make sure that you get those event though since I had to
> > twik my dsdt for that.
> > You can also look at cpufreq of acpi cpu throttling to 
> > throttle down you
> > cpu when on battery to conserve power (assuming you cpu 
> > supports that).
> 
> I took a look at laptop mode. However, as the author said "it adds a
> non-significant" amount of battery time.

Where have you read that?
Have you looked at linux/Documentation/laptop-mode.txt?


> The thing here is, even though it's flushing the buffers when there are
> others to flush, it makes not difference if the drive is still spinning and
> not in standby-mode! I can't figure out what is making the HD _not_ spin
> down.

man hdparm, hdparm -S something.

> BTW, Laptop_mode is merged into the 2.4.23-rc5 kernel (based on the
> changelogs)

As already said.

> With regard to ACPId events, can you tell me how yo get it to work? I'm
> having problems getting it running. It gets lodaded up during init, and it's
> running as root, but after logging in (GDM desktop), the events won't fire
> until I kill the acpid process and re-run it again. (once I logged in)
> I can't figure out how to get it to work. I read in one page where the
> auther said to put xhost +local somewhere at the end of /etc/X11/xinitrc.
> That didn't work either.

I don't know.  All my scripts for power managements via acpid certainly
do not use X.  I am wondering what is inside yours, but if they come
as default with a distribution, it is probably better to use another
distribution.

I doubt also the author of laptop-mode have done that.  Again, read
the documentation which come with laptop-mode.

> > I patched reiserfs to change the journal commit time 
> > according to the entry in
> > bdflush so that I can get the regular 30 sec journal flush 
> > time when on
> > ac and 10 minutes when of battery.
> > I could try making such a patch for ext3 if you want.
> 
> You mean that has to be patched at the Kernel Code Level?? Ah.. Dang!
> 

If you read the laptop-mode documentation, you will know that reiserfs
have to be modified, but not ext3.  BTW, I don't use reiserfs.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI + Laptop Power Management

[Micha Feigin]

On Fri, Nov 28, 2003 at 04:47:14PM +0100, Ducrot Bruno wrote:
> On Fri, Nov 28, 2003 at 11:21:19AM +0800, Ow Mun Heng wrote:
> > 
> > 
> > > -----Original Message-----
> > > From: Micha Feigin [mailto:michf-+lLcF8/aw9x6auLlOhE+pQ@public.gmane.org]
> > > Sent: Friday, November 21, 2003 7:13 PM
> > > To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> > > Subject: Re: [ACPI] ACPI + Laptop Power Management
> > > 
> > > 
> > > On Fri, Nov 21, 2003 at 11:12:32AM +0100, Ducrot Bruno wrote:
>                                               ^^^^^^^^^^^^
> 
> I *never* wrote the following (but there are good stuff
> inside).
> 
> > > like said here, laptopmode is what you want since it holds back dirty
> > > buffer flushes which is what you are experiencing. It will fire up the
> > > disk on reads though. I attached a script that you can use to enter
> > > laptopmode (was actually distributed with laptopmode as far as I could
> > > tell when googling for the original patch). I call it in responce to
> > > ac events from acpid to get it running when the power is disconnected.
> > > Will have to make sure that you get those event though since I had to
> > > twik my dsdt for that.
> > > You can also look at cpufreq of acpi cpu throttling to 
> > > throttle down you
> > > cpu when on battery to conserve power (assuming you cpu 
> > > supports that).
> > 
> > I took a look at laptop mode. However, as the author said "it adds a
> > non-significant" amount of battery time.
> 
> Where have you read that?
> Have you looked at linux/Documentation/laptop-mode.txt?
> 
> 
> > The thing here is, even though it's flushing the buffers when there are
> > others to flush, it makes not difference if the drive is still spinning and
> > not in standby-mode! I can't figure out what is making the HD _not_ spin
> > down.
> 
> man hdparm, hdparm -S something.
> 
> > BTW, Laptop_mode is merged into the 2.4.23-rc5 kernel (based on the
> > changelogs)
> 
> As already said.
> 
> > With regard to ACPId events, can you tell me how yo get it to work? I'm
> > having problems getting it running. It gets lodaded up during init, and it's
> > running as root, but after logging in (GDM desktop), the events won't fire
> > until I kill the acpid process and re-run it again. (once I logged in)
> > I can't figure out how to get it to work. I read in one page where the
> > auther said to put xhost +local somewhere at the end of /etc/X11/xinitrc.
> > That didn't work either.
> 
> I don't know.  All my scripts for power managements via acpid certainly
> do not use X.  I am wondering what is inside yours, but if they come
> as default with a distribution, it is probably better to use another
> distribution.
> 
> I doubt also the author of laptop-mode have done that.  Again, read
> the documentation which come with laptop-mode.
> 
> > > I patched reiserfs to change the journal commit time 
> > > according to the entry in
> > > bdflush so that I can get the regular 30 sec journal flush 
> > > time when on
> > > ac and 10 minutes when of battery.
> > > I could try making such a patch for ext3 if you want.
> > 
> > You mean that has to be patched at the Kernel Code Level?? Ah.. Dang!
> > 
> 
> If you read the laptop-mode documentation, you will know that reiserfs
> have to be modified, but not ext3.  BTW, I don't use reiserfs.
> 

I guess that is meant for me. I read that but at the time I looked at
the diff and saw that the change was done in the jfs directory, so only
when I later looked at the code to see where ext3 should be patched I
saw that that was the change for ext3.
I wrote a reply saying that, but I had some trouble with my email so I'm
not sure if it got through.

> -- 
> Ducrot Bruno
> 
> --  Which is worse:  ignorance or apathy?
> --  Don't know.  Don't care.
> 


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

RE: ACPI + Laptop Power Management

[Ow Mun Heng]

> -----Original Message-----
> From: Ducrot Bruno [mailto:ducrot-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org]
> Sent: Friday, November 28, 2003 11:47 PM
> To: Ow Mun Heng
> Cc: Micha Feigin; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> Subject: Re: [ACPI] ACPI + Laptop Power Management
> 
> 
> > I took a look at laptop mode. However, as the author said "it adds a
> > non-significant" amount of battery time.
> 
> Where have you read that?
> Have you looked at linux/Documentation/laptop-mode.txt?

I took another look at the kernel docs. It _does_ say it adds like 10% of
batt time. The initial post about the "NON-significant ..etc" was actually
gathered from one website that had the laptop-mode's author email.

 
> 
> > The thing here is, even though it's flushing the buffers 
> when there are
> > others to flush, it makes not difference if the drive is 
> still spinning and
> > not in standby-mode! I can't figure out what is making the 
> HD _not_ spin
> > down.
> 
> man hdparm, hdparm -S something.

That's just for power savings. eg: power management of the HD.. I spin the
HD down manually using hdparm -y /dev/hda




> 
> > BTW, Laptop_mode is merged into the 2.4.23-rc5 kernel (based on the
> > changelogs)
> 
> As already said.

I took a swipe at kernel 2.4.23 yesterday night. Laptop mode is in..
_however_ cpu frequency scaling is not in there and that's most certainly
something that I _need_ ton conserve batt life. I used to get it from the
-ac branch but seems like Alan is busy and heven't released anything after
2.4.22-ac4.

BTW, anyone know how to get cpufreq into the main kernel. Where can I find a
diff or is there another method of cpu freq scaling now??


> 
> > With regard to ACPId events, can you tell me how yo get it 
> to work? I'm
> > having problems getting it running. It gets lodaded up 
> during init, and it's
> > running as root, but after logging in (GDM desktop), the 
> events won't fire
> > until I kill the acpid process and re-run it again. (once I 
> logged in)
> > I can't figure out how to get it to work. I read in one 
> page where the
> > auther said to put xhost +local somewhere at the end of 
> /etc/X11/xinitrc.
> > That didn't work either.
> 
> I don't know.  All my scripts for power managements via acpid 
> certainly
> do not use X.  I am wondering what is inside yours, but if they come
> as default with a distribution, it is probably better to use another
> distribution.
> 

Fudge that.. I found out how to use xhost properly (finally) yesterday. It
seems I missed the ":" in "xhost +local:"

Now screen blanking works great and so does acpid and it's events.

 BTW, I use RedHat 9

Cheers 

OW


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/