Dell Inspiron 6000

[seppelrockt]

HowTo make the multimedia buttons and Fn + Fx keys work on Dell I6000 (by seppelrockt)

1. The multimedia buttons

This part is quite easy. With "multimedia buttons" I refer to the buttons on the front (the three volume keys in the upper right corner of the keyboard work, too).

The only think you need is the file ~/.xbindkeysrc (see below). When this file exists xbindkeys is starting automatically with X/KDE (at least in my case - Gnome users please give a feedback). Disadvantage of this approach: it only works on X, not on console. But that's OK for me.

You can download my .xbindkeysrc here.

The code is quite selfexplaining. If you wonder why each volume keys has two amixer commands that is for controling both the internal speaker and the headphone plug at the same time. It's a "workaround" cause I don't know how to detect whether a headphone is plugged (like Windows does) -- but it works [TM].

Code: Select all

#*************************************#
# Make the hotkeys on Dell I6000 work #
#*************************************#

# The keycodes of the buttons can be read with xev

## Fn keys

# Fn + F1 --> Suspend to Disk
 "sudo /usr/sbin/hibernate"
 c:139 #220

# Fn + F10 --> Eject
 "eject"
 c:209 #246

## Volume settings (for both Speaker and Headphone) 
 
# Mute/Unmute 
 "amixer set Master toggle && amixer set Headphone toggle" 
 c:160 
 
# Lower Volume 
 "amixer set Master 5%- && amixer set Headphone 5%-" 
 c:174 
 
# Raise Volume 
 "amixer set Master 5%+ && amixer set Headphone 5%+" 
 c:176 
 
## Player settings 
 
# Play/Pause 
 "amarok -t" 
 c:162 
 
# Previous Track 
 "amarok -r" 
 c:144 
 
# Next Track 
 "amarok -f" 
 c:153 
 
# Stop 
 "amarok -s" 
 c:164 

2. Hotkeys for ACPI events

I think the clearest setup would be like this:

Fn + Esc --> Suspend to RAM
Fn + F1 --> Suspend to Disk
Power button --> Power off Computer
Close LID --> Power off Display (Display must come back when LID opened)

2.1 Power button, LID open/close and Fn + Esc

This is the easier part of the ACPI events section. I assume you have Suspend-to-Disk and Suspend-to-RAM working and the following ebuilds installed: sys-power/acpid and sys-apps/vbetool.

Please check if acpid is in your runlevel, otherwise add it now:

Code: Select all

rc-update add acpid default

Finally please adjust the file /etc/acpi/default.sh like so (see below) or download mine here.

Code: Select all

#!/bin/sh
#***********************************************#
# Customized acpi script for Dell Inspiron 6000 #
#***********************************************#

set $*

group=${1/\/*/}
action=${1/*\//}

case "$group" in
        button)
                case "$action" in
                        # Power button --> Power Off
                        power)  logger "ACPI action: Power button" && \
                                sbin/init 0
                                ;;

                        # Fn + Esc --> Suspend To RAM
                        # (works out of the box via BIOS)
                        sleep)  logger "ACPI action: Sleep" && \
                                hibernate-ram &
                                ;;

                        # Turn the display back on when LID opened
                        lid)    grep -q open /proc/acpi/button/lid/LID/state && \
                                logger "ACPI action: LID OPEN" && vbetool dpms on
                                ;;

                        # All other actions
                        *)      logger "ACPI action $action is not defined"
                                ;;
                esac
                ;;

        *)
                logger "ACPI group $group / action $action is not defined"
                ;;
esac

Note: I am still writing on the Dell I6000 HowTo that covers Suspend to RAM/Disk. The setup was pretty straightforward with suspend2-sources (x86) and no patches etc. are needed. So give it a try or wait 'til the HowTo is done. Easy so far, no

2.2 The cursed Fn +F1 hotkey

Now we only have Fn + F1 --> Suspend to Disk left -- this one is realy tricky. The problem is that the hotkeys for ACPI events we have used 'til now are managed by the BIOS/kernel. But Fn + F1 is not implemented in the BIOS (thanks Dell!) and at least me can not write a kernel patch.

So we have to start from a situation like this (check your logs):

Code: Select all

atkbd.c: Unknown key pressed (translated set 2, code 0x87 on isa0060/serio0).
atkbd.c: Use 'setkeycodes e007 <keycode>' to make it known.

My approach here is to use xbindkeys to call /usr/sbin/hibernate in userspace and do not bother with ACPI events in this case.

The first step is to assign kernel level keycodes to the hotkeys that are not managed by the BIOS. I added a keycode for Fn + F2 (wireless switch) too cause I later want to hack this to get a working WLAN LED.

I put the setkeycode commands in /etc/conf.d/local.start to call them on boot automaticaly. You can download the file here

Code: Select all

# /etc/conf.d/local.start

# This is a good place to load any misc programs
# on startup (use &>/dev/null to hide output)

## Assign keycodes to Dell Inspiron 6000 hotkeys
SKC=/usr/bin/setkeycodes
# Fn + F1 (hibernate hotkey)
$SKC e00a 120
# Fn + F2 (WLAN switch)
$SKC e008 121
# Fn + F10 (eject CDROM)
$SKC e009 122

Note: If you want to use other keycodes use the command dumpkeys to find out which keycodes are not in use yet. On my I6000 I decided to use the range 120 - 124.

To test if it works run ...

Code: Select all

/etc/init.d/local restart

... and from a console outside X run

Code: Select all

showkeys

You will see that every key gives you a keypress and a keyrelease event but our Fn + F1 hotkey is missing the release event "natanoj" has kindly provided a kernel patch for this problem in the (in)famous Dell I6000 thread and I have adjusted the patch to apply cleanly against suspend2-sources-2.6.16-rX.

Download the kernel patch here

--- linux-2.6.16-suspend2-r4/drivers/input/keyboard/atkbd.c 2006-04-24 16:10:25.000000000 +0200
+++ linux-2.6.16-suspend2-r4/drivers/input/keyboard/atkbd.c 2006-05-24 00:26:12.534113096 +0200
@@ -398,6 +398,12 @@
default:
value = atkbd->release ? 0 :
(1 + (!atkbd->softrepeat && test_bit(atkbd->keycode

Code: Select all

, atkbd->dev->key)));
+			/* Workaround for Dell Inspiron 6000 laptop function keys 
+			   otherwise there is only a keypress and no keyrelease event */
+			if (code >= 0x85 && code <= 0x8b) { 
+				value = 3; 
+			} 
+			/* End of Dell Inspiron 6000 patch */
 
 			switch (value) {	/* Workaround Toshiba laptop multiple keypress */
 				case 0:
[/quote]

Note: I have to put the CODE snipets into QUOTE tags from now on cause the forum software gets totaly fooled by this patch (it contains a CODE keyword itself). Seems like the whole world is one big workaround ;-)

I assume you are familar with kernel compilation. Apply the patch from inside your kernel source dir with 
 
[quote="*** This is code too -- but I can't use the CODE tag here ***"] 
patch -p1 < /path/to/the/patch 
[/quote] 
 
and boot in your newly patched kernel. Now you should get proper keyrelease events for the hotkeys as well. But this are [i]kernel level[/i] keycodes that are translated into [i]X level[/i] keycodes by Xorg.

In my .xbindkeysrc from above the (X level) keycodes are already written down -- you can get the X keycode of a key with [i]xbindkeys -k[/i].

The last problem is that /usr/sbin/hibernate can only be run as root --even the SETUID bit didn't help here cause the script checks it permissions by itself. [url=http://www.gentoo.org/doc/en/sudo-guide.xml][i]app-admin/sudo[/i][/url] comes in handy here cause it allows us to (passwordless) perform specified actions as non-root user that are otherwise only allowed to root. sudo is much safer then the SETUID bit anyway so give it a try. 

To run the hibernate script as non-root I have added the following lines to the config file /etc/sudoers (download the config [url=http://wwwhomes.uni-bielefeld.de/sroeder/sudoers]here[/url])

[quote="*** This is code too -- but I can't use the CODE tag here ***"]
# Let all (local) users in group "wheel" suspend the maschine
%wheel          whiterabbit=    NOPASSWD: /usr/sbin/hibernate
[/quote]

whiterabbit is the [i]hostname[/i] of my maschine so you have to adjust it to your needs -- I don't know why but localhost didn't work for me (also it works in general on my system, e.g. with CUPS).

[b]3. Bonus features[/b]

Try Fn + F10 and see what happens ;-)

I didn't configure the Fn + F3 hotkey (yet) cause I don't know what to do with it ...

On my ToDo is to use xbindkeys + sudo and a script or something to get a fully functional WLAN LED that is ON when the WLAN is on (no matter if it is actually connected to a AP or not) and OFF when ... well the WLAN is OFF. This involves module loading/unloading or maybe calling some init scripts/hotplug/coldplug whatever. If [i]you[/i] have ideas how to archive this feel free to write it down here ;-)

[b]Afterword[/b]

All the download files can be found [url=http://wwwhomes.uni-bielefeld.de/sroeder/]here[/url] 

I want to thank all the people in this thread who helped to sort out all the small and bigger problems with the Dell I6000 -- which works very smoothly today with Gentoo (thanks Gentoo Devs!).

This is just a fragment of the in-the-works HowTo V2. Next "fragment" will be about Suspend modi I guess, befor I finally write down the whole thing "from the beginning" aka installing Gentoo from Scratch on the I6000.

Stay tuned ...
seppelrockt

[seppelrockt]

As editing a post doesn't trigger the "new posts in this thread" thingy I bump this manually Welcome on page 30!

[BlueShark]

God job seppelrockt!

I have two additional questions/remarks:

1) Are there specific changes on the kernel necessary, so that the "Fn+any" keys work? That is my guess. So please put your current cernel config file somewhere online.

2) I don't know why you are only turning on the lid when it is opened? Why not turnig it off, when it is closed? Did you made any additional changes, so that it turnes off by closing? When I close my lid, it stays on. So I changed your /etc/acpi/default.sh in that way:

Code: Select all

...
			# Switch the display on/off when it is opened/closed
			lid)    if [ `sed -n -e 's/state:[ ]*\(.*\)/\1/p' /proc/acpi/button/lid/LID/state` == open ]
				then
					logger "ACPI action: LID open"
					vbetool dpms on
				else
					logger "ACPI action: LID close"
					vbetool dpms off
				fi
				;;
...

Maybe your way is more elegant than this, but it works for me. But I am still interessted in your solution.

[seppelrockt]

God job seppelrockt!

Thanks

I have two additional questions/remarks:

1) Are there specific changes on the kernel necessary, so that the "Fn+any" keys work? That is my guess. So please put your current cernel config file somewhere online.

Nothing special I would say ... Maybe check the ACPI section of my kernel config.

2) I don't know why you are only turning on the lid when it is opened? Why not turnig it off, when it is closed? Did you made any additional changes, so that it turnes off by closing? When I close my lid, it stays on.

Because it turns off when closed automatically here without further tweaking - don't know what is the difference between my I6000 and yours Thanks for the feedback! I hope we can solve the remaining problems you have.

[amaroc]

Long time ago we had a discussion regarding harddisk update and whether standard drives are supported.

Regarding the HD you are right - the bridge is outside the enclosure and I would have to further disamble the laptop to remove the bridge. Then I can not simply change the HD anymore cause the SATA connector is justly loosy hanging around ih the laptop and I don't even know if it's cable is long enough.
But I read on the thinkwiki that the IBM laptop HDs that use sata-pata bridges have need a special firmware and if you do not my them directly at IBM respectivly Dell in my case they won't work. But I don't want to by a new HD from Dell, they are to expensive!
Damn, ... did anybody ever disamble the notebook any further?

I did not found any further posts regarding hdd-replacement on I6000 so I post my results here.
I've replaced the original 60GByte

Code: Select all

ATA device, with non-removable media
        Model Number:       Hitachi HTS541060G9AT00
        Serial Number:      MPB3PAX5HRLNMM
        Firmware Revision:  MB3OA61A 

by a 120GByte

Code: Select all

ATA device, with non-removable media
        Model Number:       SAMSUNG HM120JC
        Serial Number:      S0A0J10A128919
        Firmware Revision:  YL100-12

and it works fine. Performance via hdparm -tT is a liitle bit better (34 vs 30 MByte/sec) - don't know if it's really noticable - maybe X starts a little bit faster.
In addition, the Samsung is really a quite drive - you almost don't here anything from the drive. The only noise comes from the fan now .

I've simply copied via dd and "cp -a" to the new hdd via an USB-to-IDE connector. I only had to reinstall grub after installing the new hdd.
The system now runs for about 2 weeks without any problems.

[juniper]

haven't been to this thread in a while...

i noticed on the last few pages in this topic that people have been using the patchset from rtr.ca/dell to get suspend to ram working. i am not using those patches as i found that neither suspend to ram nor suspend to disk worked reliably. i am using a patched 2.6.16.16 kernel (patched with suspend 2). suspend to disk works flawlessly, but suspend to ram doesn't work at all. my computer will go to sleep and it will wake up, but there is a lot of weirdness. for example, i can't reboot or write to /tmp (i get an input/output error). i suspect i can't do anything that requires me to write to /.

anyways, has anyone gotten suspend to ram working with suspend 2?

[masteroftheuniverse]

nope, swsusp2 doesn't do suspend to ram on the i6k. you need to use the other patchset, coupled with vbetool. i've got it working with 2.6.16.4 with the patchset, without swsusp2. suspend to disk works great as well.

[juniper]

nope, swsusp2 doesn't do suspend to ram on the i6k. you need to use the other patchset, coupled with vbetool. i've got it working with 2.6.16.4 with the patchset, without swsusp2. suspend to disk works great as well.

ok, maybe i will try that patchset. the only reason why i wasn't using it was because i found BOTH suspend to ram and suspend to disk unreliable (where as now, suspend to disk works perfectly and suspend to ram works never ). i was using vbetool, so that was not the issue. maybe it was that I was using an older kernel (maybe 2.6.12 or something) so it might be better now.

do you know why/if they are fixing suspend2 on i6k (i don't really know what i6k is, but oh well)? as i stated above, my computer wakes up, but there is something wrong with / (with a whole bunch of input/output errors). i can't even shutdown (shutdown gives an input/output error).

thanks

[masteroftheuniverse]

i6k is my shorthand for inspiron 6000.
nothing worked for me either until 2.6.16. then with the patchset i was able to get suspend to ram and the sd slot to work. i have the model with the radeon gpu, so i have the hibernate script configured to use both that and vbetool.
i dunno what the status of the suspend2 patch is. it really seems like the main kernel tree's suspend function is stabilizing, tho the interface isn't as verbose.

[juniper]

@masteroftheuniverse

Note: I am still writing on the Dell I6000 HowTo that covers Suspend to RAM/Disk. The setup was pretty straightforward with suspend2-sources (x86) and no patches etc. are needed. So give it a try or wait 'til the HowTo is done. Easy so far, no

hmmmmm, I wonder if sepplerockt has gotten suspend to ram/disk working with suspend2. weird. i will give it a try with suspend2-sources. thus far i have been patching vanilla sources.

by the way, where are the patches for the 2.6.16 kernels?

http://rtr.ca/dell_i9300/

seems to be gone.

[amaroc]

by the way, where are the patches for the 2.6.16 kernels?
http://rtr.ca/dell_i9300/
seems to be gone.

Yes, it seems that rtr.ca is gone. If you want me to send you the latest patch-set you could send me your email address via pn...

Anyway, suspend-to-ram is now really stable. Stable means, I did not reboot since 28th of March w/ 2.6.16 + patchset but had at least one STR cycle every day. 2 weeks ago I've updated my harddisk and did a kernel update as well to 2.6.16.16 - sure, no reboot since.

I have never tested a blank vanilla kernel so I don't know whether STR will work reliable w/o the patchset, maybe seppelrockt had some success. You normally don't need suspend2, the built-in suspend does the job - if really still needed.
You might also find Spock's vesa-tng usefull in order to get hires framebuffer on GMA900 graphics working.

[juniper]

by the way, where are the patches for the 2.6.16 kernels?
http://rtr.ca/dell_i9300/
seems to be gone.

Yes, it seems that rtr.ca is gone. If you want me to send you the latest patch-set you could send me your email address via pn...

Anyway, suspend-to-ram is now really stable. Stable means, I did not reboot since 28th of March w/ 2.6.16 + patchset but had at least one STR cycle every day. 2 weeks ago I've updated my harddisk and did a kernel update as well to 2.6.16.16 - sure, no reboot since.

I have never tested a blank vanilla kernel so I don't know whether STR will work reliable w/o the patchset, maybe seppelrockt had some success. You normally don't need suspend2, the built-in suspend does the job - if really still needed.
You might also find Spock's vesa-tng usefull in order to get hires framebuffer on GMA900 graphics working.

sure, i could send you my address, but maybe we should put the patches in a public place? i can't be the only person looking for them.

i just tried the suspend2-sources (version 2.6.16-suspend2-r4 to be exact) and it didn't work. i used my own config file, but i doubt that was the issue. same thing as before, if i try to run a command (like vbetool) i get an input/output error.

[amaroc]

by the way, where are the patches for the 2.6.16 kernels?
http://rtr.ca/dell_i9300/
seems to be gone.

sure, i could send you my address, but maybe we should put the patches in a public place? i can't be the only person looking for them.

I've send Mark an email whether he will make the patchset accessible again.
In the meantime - I'm sorry, but I've never woked with any web-space, so I don't know where and how to make any files public Could anybody provide some space and help...

i just tried the suspend2-sources (version 2.6.16-suspend2-r4 to be exact) and it didn't work. i used my own config file, but i doubt that was the issue. same thing as before, if i try to run a command (like vbetool) i get an input/output error.

Strange - could you please post your error messages?
I assume that there is more than a "version" problem. Maybe wrong SATA configuration that comes somehow from your .config. I could provide you my working .config - but same problem as above (email would work though .

[juniper]

@amaroc

i just sent you a pm (i clicked the "pm" button on your post) hopefully you got it. also, i created a simple google webpage, we can put the patchset there so other people can access them (at least, i think we can put it there).

The error i get from my kernels is the following. after it wakes up (it takes about 20 seconds to wake up) commands seem to simply say Input/Output error. for example, if i type vbetool it will say

vbetool: Input/Output error.

also, i can't write to my / partition at all (this may cause the input/output error). i tried to touch files in /tmp and /var, but i can't. i also can't shutdown/reboot since i get the same Input/Output error.

Just to make this clear. to get STR working, people are using a patchset and the builtin suspend function of the kernel, right? this is true for everyone but seppelrockt, who appears to have STR and STD working with suspend2, right?

just clearing, a lot of dangling pronouns in the posts!uns in the posts!

well, i will try the config file that seppelrockt posted (but i don't think that is the problem). hopefully, he will get his suspend2 wiki up soon.

[amaroc]

...also, i created a simple google webpage, we can put the patchset there so other people can access them (at least, i think we can put it there).

Mark has already answered to my email (thx). There were some problems with the server that have been solved now. The link will stay in future so it's not essentially needed to shadow the patchset. Anyway, I will send you my config-file...

The error i get from my kernels is the following. after it wakes up (it takes about 20 seconds to wake up) commands seem to simply say Input/Output error. for example, if i type vbetool it will say

vbetool: Input/Output error.

also, i can't write to my / partition at all (this may cause the input/output error). i tried to touch files in /tmp and /var, but i can't. i also can't shutdown/reboot since i get the same Input/Output error.

It looks like an hdd error after resume - it's either a wrong .config or the lack of an important patch - maybe the libata_acpi_patch. There is also the chance that you have a "bad" combination of board/chipset revision and harddisk. Following some discussions here and in the kml this could also be a reason.

Just to make this clear. to get STR working, people are using a patchset and the builtin suspend function of the kernel, right? this is true for everyone but seppelrockt, who appears to have STR and STD working with suspend2, right?

Yes, most of the positive feedback is related to 2.6.16.x and Mark Lord's patch-set.
However, Mark wrote that he has his 9300 running with unmodified 2.6.16.18 and 2.6.17-rc6 kernels. I agree to his comment that sooner or later no patch will be needed anymore.

However, regarding your issues with your current configuration I propose to use gentoo's 2.6.16.16 + patchset + "working .config".

[juniper]

ok. i am compiling suspend2-sources with seppelrockt config right now... will let you know how it goes... otherwise i will probably use the patch set.

i will keep you posted, thanks for the help.

[seppelrockt]

The was quite some discussion going on here while I was away. Sorry but I am working all night this weeks to earn some money so all the linux stuff has to wait a little. Great to hear that the HD replacement works (and yeah I love Samsung HDs too for the silence and coolness).

It is always hard to answer for me how I made STR/STD working cause it more or less worked out of the box here. Your summary was correct: I use latest stable x86 suspend2-sources WITHOUT ANY patches (but the keyboard patch, but this is not related). This weekend I will try to write I short summary with my configs related to Suspend - but writing HowTos is a time consuming task in general.

Since the feedback to my last HowTo part was not so strong and nobody could help me with the remaining things I thought it would not hurt to work on this later.

[BlueShark]

Hey Guys,


how can I check if my display is on, off, stand by, suspended, etc. I googled a little and I had a look at the acpi website, but i haven't figured it out.

@seppelrockt: can you pls tell me your BIOS-Version. Maybe that is the different between your and my system.


Thx

BlueShark

[seppelrockt]

@blueshark. I have BIOS version A09. As far as I know the screen has only three "states": ON (you should see something on it), some standby mode (the screen is black but the backlight is still on) and OFF (backlight is off too).

[mhanski]

@blueshark. I have BIOS version A09. As far as I know the screen has only three "states": ON (you should see something on it), some standby mode (the screen is black but the backlight is still on) and OFF (backlight is off too).

seppelrockt: what is the highest resolution you can get on your display with A09? I slightly remember having 1680x1050 in the olde days of A05, but after subsequent bios upgrades to A07 and now to A09 I cannot get anything better than 1200x800. My bios settings say that the native resolution is 1200x800, and neither on my linux nor windows it's possible to increase this resolution. This is with ATI X300 and with both free and proprietary drivers.

Does anybody know how to change it?

[seppelrockt]

I can not help you in this case with VGA resolution cause I neither have the 1680er screen (have 1280 only hardware on the hardware side) nor do I have the ATI graphics. As far as I know the screen modes are defined in VGA BIOS (that is updated with BIOS update) and so one Intel and ATi graphics MIGHT be different.

Did anybody play with the new 2.6.17 kernel and its suspend (NOT suspend2). Is it worth switching from suspend2 to suspend?

[juniper]

has anybody tried the new xorg 7 on this laptop? specifically, with the intel 915 card?

[seppelrockt]

has anybody tried the new xorg 7 on this laptop? specifically, with the intel 915 card?

Yes I previously run xorg-7.0 and for some weeks now 7.1 with i915 - no problems so far. If you have further questions let me know ...

[BlueShark]

has anybody tried the new xorg 7 on this laptop? specifically, with the intel 915 card?

Yes I previously run xorg-7.0 and for some weeks now 7.1 with i915 - no problems so far. If you have further questions let me know ...

It also works fine with the ATI M300 card.

[juniper]

has anybody tried the new xorg 7 on this laptop? specifically, with the intel 915 card?

Yes I previously run xorg-7.0 and for some weeks now 7.1 with i915 - no problems so far. If you have further questions let me know ...

ok, I will follow the gentoo xorg 7 guide and get the thing going. I will ask if i have trouble.

also, have you seen

http://forums.gentoo.org/viewtopic-t-468064.html

?

people claim that this chip is having trouble with xorg 7.1, but i think the only issue is libGL. are you having the same problem?