Dell laptop lid switch not working

[mpc]

I have a Dell Inspiron 5100 running with the 2.6 kernal and the screen does not blank when I close the lid.

I noticed a mention of the lid switch on boot, and dmegs had this to say:

Code: Select all

ACPI: Lid Switch [LID]

Has anyone else run into this problem?

Any Ideas?

Thank you

[Earthwings]

You've got to either enable it in your BIOS or emerge acpid and configure it to react on the LID switch.

[mpc]

I have acpid emerged but I can't find a config file for it?

[Earthwings]

By default /etc/acpi/default.sh is called for every event.

Here's a little more extended one. Replace ACTION with what you like.

Code: Select all

#!/bin/sh

set $*

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

case "$group" in
	button)
		case "$action" in
			power)	ACTION
				;;
				
			sleep)  ACTION
				;;
			
			lid)	ACTION
				;;

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

	battery)
		case "$action" in
			battery) ACTION 
				 ;;
			*)	 logger "ACPI group battery / action $action is not defined"
				 ;;
		esac
		;;
				
	ac_adapter)
		case "$action" in
			ac_adapter) ACTION
				    ;;
			*)	    logger "ACPI group ac_adapter / action $action is not defined"
				    ;;
		esac
		;;
	*)
		logger "ACPI group $group / action $action is not defined"
		;;
esac

See [1] for an example on how to use for Power Management.

[1] http://www.stud.uni-karlsruhe.de/~uxhz/ ... management

[mpc]

Thanks for your help. Sory about the late reply. Work called and all of that.

Sorry if I'm asking stupid questions but I can't seam to find a way to get the screen to blank. So therefor I can not modify the default.sh script

I tried:

Code: Select all

setterm -powersave powerdown

but all I got was:

Code: Select all

cannot (un)set powersave mode

I know I'm missing somthing realy simple. Sorry.

Thanks for all of your help

[Earthwings]

Try

Code: Select all

xset dpms force off

. Replace off with suspend or standby if you like.

[mpc]

tried

xset dpms force off

the system takes the command but nothing happens

[Earthwings]

Do you have

Code: Select all

Option  "DPMS"  "true"

in the Monitor section in /etc/X11/XF86Config?

[mpc]

AHHH!!! Wonderfull. By adding:

Code: Select all

Option  "DPMS"  "true"

I was able to get the screen to blank. But...

1) The screen blanks but it looks like the backlight is still on. Any Idea how to kill this?

2) And second, I added the lid line to /etc/aspci/default.sh as follows with no result.

Code: Select all

case "$action" in
			power)	/sbin/init 0
				;;
			lid)	/usr/X11R6/bin/xset dpms force off
				;;
			*)	logger "ACPI action $action is not defined"
				;;
		esac

Thank you very much for all of your help.

[Earthwings]

You can use app-laptop radeontool if you have an ATI Radeon mobility graphics card to turn the backlight off.

For the script: Very likely there's a syntax error inside that causes the script to fail. Try running it by hand and look for useful error messages.

[zerocool_australia]

Dell Inspiron 1100 and 5100 series laptops' BIOS does not work with the Linux ACPI implementation currently, the biggest problem is the /proc/acpi/events does not generate any events (from what I can see) ever. I have had both an 1100 and a 5100 and neither will do this. You can get the status of everything from /proc/acpi/[device] but no realtime event generation.
DPMS in X will blank the screen and or power down the backlight also when you have DPMS on. If it is not turning off the backlight it probably isn't actually DPMS doing it. Mess with the xset dpms command sometime. I put an xset dpms 300 300 300 in my flux rc file to turn off the screen totally after 5 minutes and it works. All those acpid scripts you can throw out the window, acpid is dependant on /proc/acpi/events working

David

[mpc]

All those acpid scripts you can throw out the window, acpid is dependant on /proc/acpi/events working

That explains a lot

However seeing as

Code: Select all

xset dpms force off

still only blanks the screen and does not kill the backlight. I looks as if dpms is enabled implicity when dpms force off is entered. Testing dpms with

Code: Select all

xset dpms 10 10 10

just put the screen blank but did not kill the backlight. I must me missing somthing

radeontool kills the backlight but I had to turn it back on headless (not somthing I want to do all the time)

[zerocool_australia]

Very interesting, for me, xset dpms force off turns off the backlight.
Here is my xset -q, paste yours too please possibly there's another setting I forgot that I set.

bash-2.05b$ xset q
Keyboard Control:
auto repeat: on key click percent: 0 LED mask: 00000000
auto repeat delay: 660 repeat rate: 25
auto repeating keys: 00ffffffdffffbbf
fadfffdfffdfe5ef
ffffffffffffffff
ffffffffffffffff
bell percent: 50 bell pitch: 400 bell duration: 100
Pointer Control:
acceleration: 2/1 threshold: 4
Screen Saver:
prefer blanking: yes allow exposures: yes
timeout: 600 cycle: 600
Colors:
default colormap: 0x20 BlackPixel: 0 WhitePixel: 16777215
Font Path:
/usr/X11R6/lib/X11/fonts/misc/,/usr/X11R6/lib/X11/fonts/75dpi/,/usr/X11R6/lib/X1 1/fonts/100dpi/
Bug Mode: compatibility mode is disabled
DPMS (Energy Star):
Standby: 300 Suspend: 300 Off: 300
DPMS is Enabled
Monitor is On
Font cache:
hi-mark (KB): 5120 low-mark (KB): 3840 balance (%): 70

File paths:
Config file: /etc/X11/XF86Config
Modules path: /usr/X11R6/lib/modules
Log file: /var/log/XFree86.0.log

[mpc]

Here you go

Code: Select all

Keyboard Control:
  auto repeat:  on    key click percent:  0    LED mask:  00000000
  auto repeat delay:  660    repeat rate:  25
  auto repeating keys:  00ffffffdffffbbf
                        fadfffffffdffdff
                        ffffffffffffffff
                        ffffffffffffffff
  bell percent:  50    bell pitch:  400    bell duration:  100
Pointer Control:
  acceleration:  2/1    threshold:  4
Screen Saver:
  prefer blanking:  yes    allow exposures:  yes
  timeout:  0    cycle:  0
Colors:
  default colormap:  0x20    BlackPixel:  0    WhitePixel:  16777215
Font Path:
  /usr/X11R6/lib/X11/fonts/local/,/usr/X11R6/lib/X11/fonts/misc/,/usr/X11R6/lib/X11/fonts/75dpi/:unscaled,/usr/X11R6/lib/X11/fonts/100dpi/:unscaled,/usr/X11R6/lib/X11/fonts/Speedo/,/usr/X11R6/lib/X11/fonts/Type1/,/usr/X11R6/lib/X11/fonts/75dpi/,/usr/X11R6/lib/X11/fonts/100dpi/
Bug Mode: compatibility mode is disabled
DPMS (Energy Star):
  Standby: 7200    Suspend: 7200    Off: 14400
  DPMS is Enabled
  Monitor is On
Font cache:
  hi-mark (KB): 5120  low-mark (KB): 3840  balance (%): 70
File paths:
  Config file:  /etc/X11/XF86Config
  Modules path: /usr/X11R6/lib/modules
  Log file:     /var/log/XFree86.0.log

I did notice that dpms was not enabled when I turned my computer back on, but I enabled it and xset dpms force off still will not kill the backlight

Thanks

[kongit]

Dell Inspiron 1100 and 5100 series laptops' BIOS does not work with the Linux ACPI implementation currently, the biggest problem is the /proc/acpi/events does not generate any events (from what I can see) ever. I have had both an 1100 and a 5100 and neither will do this. You can get the status of everything from /proc/acpi/[device] but no realtime event generation.
DPMS in X will blank the screen and or power down the backlight also when you have DPMS on. If it is not turning off the backlight it probably isn't actually DPMS doing it. Mess with the xset dpms command sometime. I put an xset dpms 300 300 300 in my flux rc file to turn off the screen totally after 5 minutes and it works. All those acpid scripts you can throw out the window, acpid is dependant on /proc/acpi/events working

not entirely true. I have a dell inspiron 5100 and I know that the lids states are changed in the file. It might be because your bios isn't a newer one. I had to upgrade mine to at least a23 to get acpi to work. However I have not been able to get the screen to turn off when shutting the lid but that is most likely due to some error on my part.

[trapperjohn]

I have an Inspiron 8600 and lid etc. works. I've read somewhere, that newer nVidia Drivers don't work with DPMS - maybe there is a problem with your drivers, too. ATM I have to restart acpid after login, else "xset" doesn't work - still have to investigate this problem.

Another thing is replacing your DSDT with a corrected one - look at http://acpi.sf.net (there's a HowTo in this forum, too).

Good Luck.

[zerocool_australia]

not entirely true. I have a dell inspiron 5100 and I know that the lids states are changed in the file. It might be because your bios isn't a newer one. I had to upgrade mine to at least a23 to get acpi to work. However I have not been able to get the screen to turn off when shutting the lid but that is most likely due to some error on my part.

Yes, sigh, I knew that, my point is that the /proc/acpi/event interface does not work and you will not get realtime events. If you want to do anything with the status of the lid, you would have to have a daemon poll the lid state file. I never said this didn't work.

By the way, the newer BIOSes are the reasons this don't work. For lucky people with A02 and other early BIOSes more ACPI seems to work.


Also....as someone else mentioned new DSDTs...there is no new DSDT to fix the event interface and I don't have enough time to write one.

[g3n]

I have bios a23, is there anything that i can do? i cant understand this

[_DaMaGe_]

Hey guys, I had a similar problem
I just bought a dell inspiron 1150 and the lid switch wasn't working
i don't think any of the events were working
I'm pretty sure the ACPID daemon just isn't compatible with these dells
however, querying the states of the devices seemed to be working fine, as my battery monitor works fine
so i created a perl script to run as a daemon to monitor the lid switch and turn it off with the xset dpms force off command
here is the script, it may work for you too

<CODE>
#!/usr/bin/perl -w
$pid = fork();
if($pid) { exit; }
$last = 1;
while(1) {
sleep 3;
@data = `cat /proc/acpi/button/lid/LID/state`;
$state = $data[0];
if($state =~ /open/ && $last == 0) {
`xset dpms force on`;
$last = 1;
} elsif ($state =~ /closed/ && $last == 1) {
`xset dpms force off`;
$last = 0;
}
}
</CODE>

[DaMaGe]

Get your own name. Stop copying other people

Had this name for 7 years on the net. And slowly I start seeing other people using the SAME way of me spelling it.

[_DaMaGe_]

to be truly honest, I never even noticed your nickname
I've been nicknamed damage for most my life because it rhymes with my last name

sorry about that!

i've been trying to think of a new nickname anyways obviously its a decent nickname for my friends to call me , but not on the net where we are running out of 'unique' handles.

[_DaMaGe_]

well, anyways
i'll use this nickname for now
can't go wrong with japanese...

[_DaMaGe_]

I tried to change it... doesn't seem to work
sorry!

[DaMaGe]

lol, not everyday you come face to face with yourself!!!!!

[Voltago]

Had the same issue with the lid some time ago.
What bugged me was that you seem to need a running X to use the xset tool to control monitor states, also trying it in terminals other than vc 7 where X was running didn't work.
So I figured out how X was doing it and hacked up a tool that does exactly the same directly. Actually, I just altered the 'vbetest' tool a bit, see code comment. OK, here goes:

Code: Select all

/*
Change the dpms mode of your graphics controller.

Version: 0.1 
Date: 12.06.2004

This program is derived from the public domain 
code of 'vbetest.c' from the 'lrmi' package 
available at http://sourceforge.net/projects/lrmi 
and is also in the public domain.
Furthermore, this program comes with no warranty for
fitness for any purpose whatsoever and if it reduces
your computer to a smoldering heap I shall not be
held responsible.

To build it, you'll need said lrmi package.
I used version 0.8, the environment is Linux,
the machine is a Dell 510m laptop with Intel's
82852/855GM graphics bloody controller.

Andreas Eckstein, 2004
*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/ioctl.h>
#if defined(__linux__)
#include <sys/io.h>
#include <sys/kd.h>
#include <sys/stat.h>
#elif defined(__NetBSD__)
#include <time.h>
#include <dev/wscons/wsconsio.h>
#include <machine/sysarch.h>
#elif defined(__FreeBSD__)
#include <machine/console.h>
#include <machine/sysarch.h>
#endif

#include <lrmi.h>
#include <vbe.h>


#define DPMS_ON		((unsigned int)0x00000000)
#define DPMS_STANDBY	((unsigned int)0x00000100)
#define DPMS_SUSPEND	((unsigned int)0x00000200)
#define DPMS_OFF	((unsigned int)0x00000400)

#define VESA_DPMS ((unsigned int)0x00004F10)

#define DPMS_SET_MODE ((unsigned int)0x00000001)
#define DPMS_GET_MODE ((unsigned int)0x00000002)

#define MASK_LBYTE ((unsigned int)0x000000FF)
#define MASK_HBYTE ((unsigned int)0x0000FF00)

struct {
	struct vbe_info_block *info;
	struct vbe_mode_info_block *mode;
	char *win; 	/* this doesn't point directly at the window, see update_window() */
	int win_low, win_high;
} vbe;

void
usage_and_quit(int error)
{
	fputs("Usage: dpmsctl {on|standby|suspend|off|status}\n",
	 error ? stderr : stdout);
	exit(error);
}

int
main(int argc, char *argv[])
{
	int retval=0;
	struct LRMI_regs r;
#if defined(__NetBSD__)
	unsigned long iomap[32];
#endif
	unsigned int dpms_mode=-1;
	unsigned int dpms_action=DPMS_SET_MODE;
		
	if(argc>1) {
		if (strcmp(argv[1], "on") == 0) {
			dpms_mode=DPMS_ON;
		} else if (strcmp(argv[1], "standby") == 0) {
			dpms_mode=DPMS_STANDBY;
		} else if (strcmp(argv[1], "suspend") == 0) {
			dpms_mode=DPMS_SUSPEND;
		} else if (strcmp(argv[1], "off") == 0) {
			dpms_mode=DPMS_OFF;
		} else if (strcmp(argv[1], "status") == 0) {
			dpms_action=DPMS_GET_MODE;
			dpms_mode=0;
		} else {
			usage_and_quit(1);
		} 
	}
	
	if(dpms_mode==-1)
		usage_and_quit(1);

	//printf("dpms_mode: %d\n", dpms_mode);
	
	if (!LRMI_init()){
		fprintf(stderr, "Can't initialize LRMI environment\n");
		return 1;
	}
	
	vbe.info = LRMI_alloc_real(sizeof(struct vbe_info_block)
	 + sizeof(struct vbe_mode_info_block));

	if (vbe.info == NULL) {
		fprintf(stderr, "Can't alloc real mode memory\n");
		return 1;
	}

	vbe.mode = (struct vbe_mode_info_block *)(vbe.info + 1);

#if 0
	/*
	 Allow read/write to video IO ports
	*/
	ioperm(0x2b0, 0x2df - 0x2b0, 1);
	ioperm(0x3b0, 0x3df - 0x3b0, 1);
#else
	/*
	 Allow read/write to ALL io ports
	*/
#if defined(__linux__)
	ioperm(0, 1024, 1);
	iopl(3);
#elif defined(__NetBSD__)
	memset(&iomap[0], 0xff, sizeof(iomap));
	i386_set_ioperm(iomap);

	i386_iopl(3);
#elif defined(__FreeBSD__)
	i386_set_ioperm(0, 0x10000, 1);
#endif
#endif

	memset(&r, 0, sizeof(r));

	r.eax = 0x4f00;
	r.es = (unsigned int)vbe.info >> 4;
	r.edi = 0;

	memcpy(vbe.info->vbe_signature, "VBE2", 4);

	if (!LRMI_int(0x10, &r)) {
		fprintf(stderr, "Can't get VESA info (vm86 failure)\n");
		return 1;
	}

	if ((r.eax & 0xffff) != 0x4f || strncmp(vbe.info->vbe_signature, "VESA", 4) != 0) {
		fprintf(stderr, "No VESA bios\n");
		return 1;
	}
	
	r.eax=VESA_DPMS;
	r.ebx= dpms_action | dpms_mode;
	LRMI_int(0x10, &r);
	
	if(dpms_action==DPMS_SET_MODE)
	{
		
	}
	
	if(dpms_action==DPMS_GET_MODE)
	{
		dpms_mode = (r.ebx & MASK_HBYTE);
		
		printf("active dpms mode: ");
		switch(dpms_mode)
		{
			case DPMS_ON:
				printf("on\n");
				break;
			case DPMS_STANDBY: 
				printf("standby\n");
				break;
			case DPMS_SUSPEND:
				printf("suspend\n");
				break;
			case DPMS_OFF:
				printf("off\n");
				break;
			default:
				printf("error: received non VESA-compliant DPMS state\n");
				printf("but if you can read this, DPMS state should be 'on'\n");
				retval = 1;
		}		
	}
	
	int dpms_return = (r.eax & MASK_HBYTE) >> 8;
	
	if(dpms_return!=0)
		printf("error code: %d\n", dpms_return);
	
	LRMI_free_real(vbe.info);

	return retval;
}

To build the code, save it as e. g. 'dpmsctl.c', emerge 'lrmi' and type

Code: Select all

gcc -o dpmsctl dpmsctl.c -llrmi

Oh yes, lrmi-0.8 is not in portage, but you can just rename the 0.7 ebuild.

This should make the perfect tool for use with powermanagement scripts and for driving people mad with mysterious monitor 'malfunctions' you can control per ssh...

Questions, comments and constructive criticism are most welcome!

P. S.: Look, ma, my first piece of OSS!