Xorg exits when it receives an ACPI button/lid event

Forrest: Do you think this is a duplicate of bug 61175?
https://launchpad.net/bugs/61175

No, because it happens on the livecd, and I don't have a new ATI card.

/var/log/acpid:
## Closed lid
[Thu Jul 10 00:50:25 2003] received event "video VID 00000080 00000000"
[Thu Jul 10 00:50:25 2003] notifying client 5838[105:109]
[Thu Jul 10 00:50:25 2003] notifying client 5673[0:0]
[Thu Jul 10 00:50:25 2003] client has disconnected
[Thu Jul 10 00:50:25 2003] notifying client 5673[0:0]
[Thu Jul 10 00:50:25 2003] completed event "video VID 00000080 00000000"
[Thu Jul 10 00:50:25 2003] received event "button/lid LID 00000080 00000003"
[Thu Jul 10 00:50:25 2003] notifying client 5838[105:109]
[Thu Jul 10 00:50:25 2003] notifying client 5673[0:0]
[Thu Jul 10 00:50:25 2003] executing action "/etc/acpi/lid.sh"
[Thu Jul 10 00:50:25 2003] BEGIN HANDLER MESSAGES
[Thu Jul 10 00:50:25 2003] END HANDLER MESSAGES
[Thu Jul 10 00:50:25 2003] action exited with status 0
[Thu Jul 10 00:50:25 2003] completed event "button/lid LID 00000080 00000003"
## Opened lid
[Thu Jul 10 00:50:27 2003] received event "button/lid LID 00000080 00000004"
[Thu Jul 10 00:50:27 2003] notifying client 5838[105:109]
[Thu Jul 10 00:50:27 2003] notifying client 5673[0:0]
[Thu Jul 10 00:50:27 2003] client has disconnected
[Thu Jul 10 00:50:27 2003] executing action "/etc/acpi/lid.sh"
[Thu Jul 10 00:50:27 2003] BEGIN HANDLER MESSAGES
[Thu Jul 10 00:50:27 2003] END HANDLER MESSAGES
[Thu Jul 10 00:50:27 2003] action exited with status 0
[Thu Jul 10 00:50:27 2003] completed event "button/lid LID 00000080 00000004"
[Thu Jul 10 00:50:27 2003] received event "video VID 00000080 00000000"
[Thu Jul 10 00:50:27 2003] notifying client 5838[105:109]
[Thu Jul 10 00:50:27 2003] completed event "video VID 00000080 00000000"

/var/log/messages:
## Closed lid
Jul 10 00:50:25 ubuntu kernel: [17181486.992000] video bus notify
Jul 10 00:50:25 ubuntu gnome-power-manager: (ubuntu) DPMS blanking screen because the lid has been closed on ac power
## Opened lid
Jul 10 00:50:27 ubuntu kernel: [17181488.516000] video bus notify
## Me logging in, and starting a new session
Jul 10 00:50:38 ubuntu gconfd (ubuntu-5758): Received signal 15, shutting down cleanly
Jul 10 00:50:38 ubuntu gconfd (ubuntu-5758): Exiting
Jul 10 00:51:04 ubuntu gconfd (ubuntu-7047): starting (version 2.14.0), pid 7047 user 'ubuntu'

I'm seeing this on a Dell Inspiron 4100 with an ATI Mobility M6.

The backtrace from the Xorg.0.log is as follows:

ProcXCloseDevice to close or not ?

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x81) [0x80c38e1]
1: [0xffffe420]
2: /lib/tls/i686/cmov/libc.so.6(__strtoul_internal+0x3f) [0xb7d9c7cf]
3: /usr/X11R6/bin/X [0x80e5997]
4: /usr/X11R6/bin/X(xf86HandlePMEvents+0x3b) [0x80d25bb]
5: /usr/X11R6/bin/X(xf86Wakeup+0x153) [0x80c4d93]
6: /usr/X11R6/bin/X(WakeupHandler+0x59) [0x808a599]
7: /usr/X11R6/bin/X(WaitForSomething+0x1b9) [0x81943d9]
8: /usr/X11R6/bin/X(Dispatch+0x82) [0x8086802]
9: /usr/X11R6/bin/X(main+0x485) [0x806e715]
10: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc) [0xb7d848cc]
11: /usr/X11R6/bin/X(FontFileCompleteXLFD+0xa1) [0x806da51]

Fatal server error:
Caught signal 11. Server aborting

Killing acpid is the only workaround I've found.

I have a Dell Lattitude C600 with an ATI Mobility card, also.
Killing acpid worked, but that obviously isn't ideal.

The only debugging I have done is that Xorg listens to acpid on a socket, and exits when it gets either the 'video' or 'button/lid' event.

I have seen the same issue today on an ASUS A2D with a Radeon Mobility 9200.

Notice that removing the battery had the same effect (power cable was plugged, of course...) : kill X.

Isn't the X server a little touchy about acpid events ?

I have some new things to say about the problem :
(1) I'm using the radeon driver, not fglrx ;
(2) the problem occurs with 2.6.17-10.28 and 2.6.17-10.29
(3) but not with my trusty 2.6.15-27.48 (I upgraded from dapper recently)

What information could I provide to help with finding a solution?

How do we confirm this?

(1) enough reports ;
(2) enough information to solve it.

Confirming bug as it is reported by several users. Marking Importance as High as this seems to be regression from Dapper for several users with different ATI cards.

It looks like that Xorg's power managements kicks in. Can you try disabling Xorg's PM by adding the following to you /etc/X11/xorg.conf:

Section "ServerFlags"
   Option "NoPM"
EndSection

Note: if you already have "ServerFlags" section, just add Option "NoPM" there.

This may workaround the problem for you.

Notice my comment about successfully using an older kernel with the rest being the same.

I'm not sure playing with the xserver will really help :-/

Adding that section to xorg.conf does work around the problem. Hopefully that provides enough information for someone to figure out and fix this problem.

See: http://www.linuxforums.org/forum/linux-laptops/58492-dell-c610-display-close-crashes-x-org.html

It claims: "The problem appears to be solved. Disabling the console framebuffer fixed it."

Imho i dont think it's a xorg related problem I dont use fbdev under xorg and I tried the "nopm" option but it doesnt work for me.
And today I noticed that X crashes just pressing the lock-screen button very very very boryng, I think it's a kernel related bug...

I'm experiencing the same issue on a Dell Latitude C640. The workaround works, sort of. The X-server doesn't crash but the DPMS remains off. Only by remotely accessing the macheine, issuing a xset dpms force on, makes the machine come alive again. (Short of the three-finger-salute)

Its not related to ATI at all... on my Dell Inspiron 8100 with a NVIDIA GeForce 2 Go, I have this same exact bug, with both the proprietary nvidia and open source nv driver. Its not card/driver related obviously.

As I said, the fact that booting a different kernel makes the difference between "works" and "fails" is certainly indicative of a bug between acpid and kernel.

Could others test if using an older kernel helps ?

I get the same problem on an old dell latitude CPx with a neomagic card. Worked perfectly under dapper.

Proposed fix from https://bugs.freedesktop.org/show_bug.cgi?id=8776:
https://bugs.freedesktop.org/attachment.cgi?id=7531

It seems to be more of a workaround than fixing the actual problem.

Hmmm... if it's handling a condition which wasn't handled, then it's a real fix :-)

If it's handling a condition which shouldn't have arisen in the first place, then it's a workaround and it's bad :-)

Why'd it get rejected?

It wasn't really rejected : it was marked as duplicate. Probably this bug should track the other bug upstream.

I experienced this bug too on a dell inspiron 8200 laptop with nvidia or nv driver. I "fixed" it by passing -noacpi to the X server command line which stops the X server from monitoring the acpi events (and the lid close). Things work fine now, however this is just a workaround until the real culprit is found. Sounds like the freedesktop bug mentioned might be the a likely source?

I am running Kubuntu 6.10, having just upgraded from 6.06.1 -- this problem appeared for me after the upgrade. I'm running Linux 2.6.17-10-generic and Xorg 7.1.1ubuntu6 and currently using nvidia-glx-legacy 1.0.7184+2.6.17.5-11. Now when I close the lid, X crashes with the following in /var/log/kdm.log.

---
Fatal server error:
Caught signal 11. Server aborting
Backtrace:
0: /usr/bin/X(xf86SigHandler+0x81) [0x80c3971]
1: [0xffffe420]
2: /lib/tls/i686/cmov/libc.so.6(__strtoul_internal+0x3f) [0xb7d577df]
3: /usr/bin/X [0x80e5a27]
4: /usr/bin/X(xf86HandlePMEvents+0x3b) [0x80d264b]
5: /usr/bin/X(xf86Wakeup+0x153) [0x80c4e23]
6: /usr/bin/X(WakeupHandler+0x59) [0x808a5c9]
7: /usr/bin/X(WaitForSomething+0x1b9) [0x8194489]
8: /usr/bin/X(Dispatch+0x82) [0x8086832]
9: /usr/bin/X(main+0x485) [0x806e715]
10: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc) [0xb7d3f8cc]
11: /usr/bin/X(FontFileCompleteXLFD+0xa1) [0x806da51]

Fatal server error:
Caught signal 11. Server aborting
---

My laptop is a Dell Inspiron 8100 with the nVidia GeForce2Go, so this is NOT an ATI problem. It is a problem that appeared between the latest version of Dapper (I always run the updates as they appear) and Edgy as of the install time, from the DVD release.

I had the same problem with a dell latitude C640 with
ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500]

X crashed when I closed the screen. This happened under Dapper with xorg-air but not with the default x server. From Edguy this happened everytime, but since I added the :
Section "ServerFlags"
   Option "NoPM"
EndSection
the problem doesn't appear anymore

/var/log/messages :
Nov 6 02:06:16 latitude gnome-power-manager: (kjo) Extinction de l'écran (DPMS) car L'écran du portable a été rabattu (branché sur secteur)
Nov 6 02:06:16 latitude kernel: [ 313.599000] video bus notify
Nov 6 02:06:16 latitude gconfd (kjo-5081): Réception du signal 15, arrêt correct
Nov 6 02:06:16 latitude gconfd (kjo-5081): Sortie
Nov 6 02:06:18 latitude kernel: [ 316.327000] mtrr: 0xe0000000,0x8000000 overlaps existing 0xe0000000,0x2000000
Nov 6 02:06:18 latitude last message repeated 2 times
Nov 6 02:06:18 latitude kernel: [ 316.327000] agpgart: Found an AGP 2.0 compliant device at 0000:00:00.0.
Nov 6 02:06:18 latitude kernel: [ 316.327000] agpgart: Putting AGP V2 device at 0000:00:00.0 into 1x mode
Nov 6 02:06:18 latitude kernel: [ 316.328000] agpgart: Putting AGP V2 device at 0000:01:00.0 into 1x mode
Nov 6 02:06:18 latitude kernel: [ 316.544000] [drm] Setting GART location based on new memory map
Nov 6 02:06:18 latitude kernel: [ 316.544000] [drm] writeback test succeeded in 1 usecs

A proper fix has been released in the xorg bugzilla.
Attaching it to this bug!

So, now that a fix is released, can we expect a package to come down our beloved broadband lines?

Nico

good question, I havent got any updates on Xorg packages the last 2 days. Anyone know where there is any info about how bugfixerelese process works?

Now, that's odd. The bug is confirmed, the importance is marked "high", a proper fix has been released (which works, btw) -- yet there are no updated packages.

I got little experience in packaging, but the deb-package in the attachment works for me. All I did was apply the fix posted above manually and rebuild the package.

Its fixed upstream not yet uploaded for ubuntu. That may change soon. also make sure the edgy-updates repo is added to sources.list or enabled if already there, as it may come down that repo. Im not working on this project but those are the main reasons why you dont see an update right now.

or dapper-updates repos if on dapper (not sure if this was a problem in dapper or just edgy.

It is not in dapper, just edgy.

Unfortunately that package causes the adept-updater to claim that it has a
newer package.

Fortunately, however, if you ignore that, the fix DOES work!

Found it on a Dell Latitude ATI Mobility M7 as others have mentioned. Just attaching to this thread to I get updates. Thanks all! I appreciate doing a google search and finding that others have grabbed the bull by the horn. This is a significant problem. Just speaking from a pragmatic standpoint (stepping back from HOW to properly fix it), my laptop is basically unusable as a laptop. I can use it as a desktop, but one which I can't easily shut off the monitor.

Thanks again,
Matt

will the .deb package work for kubuntu as well? Also, the .patch file, is that supposed to be a script, because when I run it in konsole, it errors out.

Thanks,
Ryan

Works fine -- all Ubuntu packages are interchangable. The patch is not
a script, it is a context diff to be used with 'patch'. Search Google
for more info.

On 11/25/06, GitarGr8 <email address hidden> wrote:
> will the .deb package work for kubuntu as well? Also, the .patch file,
> is that supposed to be a script, because when I run it in konsole, it
> errors out.
>
> Thanks,
> Ryan
>
> --
> Xorg exits when it receives an ACPI button/lid event
> https://launchpad.net/bugs/61746
>

Hi, just marked my bug #73819 as duplicate. The here posted deb works for me and PLEASE make this update HIGH PRIORITY as the update manager always tries to update the fixed package :(

BTW: the latest manual patch is https://bugs.freedesktop.org/attachment.cgi?id=7553

yes, here posted deb works for me too. this is a critical update, deploy it soon please!

Apparently a new version of the xserver has been released. Awkwardly enough, it still doesn't contain the fix posted in this thread. Figuring it's more than just me who needs a fixed package, I thought I might as well upload the .deb Package I just built from the latest Ubuntu-sources and the fix above.

It works for me, your mileage may vary...

I wonder why they don't include the patch -- whoever "they" may be... at least for me a standby-incapable xserver renders my Dell C600 completely unusable...

As can be easily seen from the posts in this thread, this bug affects Edgy.

I would tend to agree with Daniel.

We have a bug with an agreed upon severity with an easy fix, with low regression risks.

Then we have an update of the package released, without this fix!

What is the process involved here?

Can't there be a bug fix with a security update? Do we have to wait for a backport or another update type?

***************

Daniel, your attachmant is the -dev package. At least that is what the name implies. Is this right?

Thanks,

Nico

Nico,

thanks for pointing out this mistake. Obviously it's the wrong package :-) You'll find the right one below.
Is there a way I can delete the one above or at least edit the respective post?

Daniel

I have the same bug (ASUS A2500D, ATi mobility radeon 9600).

Your package have solved this problem, but it seems to me that there is again a new version of xserver-xorg-core. At least apt-get update always would like to update (downgrade) my xserver.

Isn't it possible that the freedesktop guys forgot about this bug?

Viktor,

the reason why apt, update-manager and so on try to replace the fixed package is that they realize that the replacement package doesn't come from the distribution's repositories. Even worse -- if you enable the automatic installation of security updates without your confirmation, guess what happens... Every now and then you find yourself with a frozen laptop, because your xserver-xorg-core got replaced without you being aware of it :-/

Let's hope it's fixed in Feisty Fawn...

Daniel

I think that it should be at least backported to Edgy, although
according to my understanding such degradings are considered to be
general updates. I was wondering if this patch was incorporated in the
xorg core source, and if I have understood well the comments, then the
source had to be patched to get the deb file.

V

On 2/3/07, Daniel Rheinbay <email address hidden> wrote:
> Viktor,
>
> the reason why apt, update-manager and so on try to replace the fixed
> package is that they realize that the replacement package doesn't come
> from the distribution's repositories. Even worse -- if you enable the
> automatic installation of security updates without your confirmation,
> guess what happens... Every now and then you find yourself with a frozen
> laptop, because your xserver-xorg-core got replaced without you being
> aware of it :-/
>
> Let's hope it's fixed in Feisty Fawn...
>
>
> Daniel
>
> --
> Xorg exits when it receives an ACPI button/lid event
> https://launchpad.net/bugs/61746
>

This issue is still present in Feisty.
I've contacted Sebastian Bacher, notifying him about the importance of this bug.

OK folks,

I got tired of having to be careful not to have my xserver-xorg-core package upgraded, and dug a bit around package building.

Apparently adding an entry in the changelog file with a new version number should be good, after having applied the patch, of course.

Now, in order not to break legitimate updates (security or backport, and most importantly dist-upgrades at a later time), I need to learn more about the naming conventions for that sort of stuff.

Where is the place to learn about this?

Dear lazyweb... Could someone even give me a ready-made version?

Would this be sufficient?

xorg-server (1:1.1.1-0ubuntu12.1+acpifix) edgy-security; urgency=low

  * Fix ACPI Lid bug - lauchpad #61746

 -- Nicolas Will <email address hidden> Fri, 9 Feb 2007 18:15:41 -0600

Thanks for the help.

Once I have a .deb that works fine and once I am certain I will not be the cause of failed dist-upgrades in the future, I will be more than happy to share, of course.

Nico

OK

I have a patched (using debian/patches, even!) package with a version number higher than the original, with no update triggers.

Compiling anything on a 1 GHz P3 Coppermine and a 4200RPM HDD is fun !

Package name is the following:

xserver-xorg-core_1.1.1-0ubuntu12.1+acpifix_i386.deb

I haven't read the MOTU wiki yet, so this will probably break everything...

More on this later, once I have figured it out.

Nico

I made another package using a proper "dch -i" for the changelog generation. It gave me:

xserver-xorg-core_1.1.1-0ubuntu12.2_i386.deb

When I installed it after

xserver-xorg-core_1.1.1-0ubuntu12.1+acpifix_i386.deb

there was nothing said about downgrading the version.

So the +acpifix should be still a lower version than the next official iteration. This has been confirmed by reinstalling it, with downgrade message and all.

So, here is the package, it should not break anything and it will be clearer about the exact version and what happened to the package.

Nico

Thank you much for the work on this! I personally am not running Linux right
now (hardware failure and am using a backup machine with Windows), but I
remember how annoying this one was, and I will have a new Latitude soon.

Thanks!

On 2/9/07, Nico <email address hidden> wrote:
>
> OK folks,
>
> I got tired of having to be careful not to have my xserver-xorg-core
> package upgraded, and dug a bit around package building.
>
> Apparently adding an entry in the changelog file with a new version
> number should be good, after having applied the patch, of course.
>
> Now, in order not to break legitimate updates (security or backport, and
> most importantly dist-upgrades at a later time), I need to learn more
> about the naming conventions for that sort of stuff.
>
> Where is the place to learn about this?
>
> Dear lazyweb... Could someone even give me a ready-made version?
>
> Would this be sufficient?
>
> xorg-server (1:1.1.1-0ubuntu12.1+acpifix) edgy-security; urgency=low
>
> * Fix ACPI Lid bug - lauchpad #61746
>
> -- Nicolas Will <email address hidden> Fri, 9 Feb 2007 18:15:41
> -0600
>
> Thanks for the help.
>
> Once I have a .deb that works fine and once I am certain I will not be
> the cause of failed dist-upgrades in the future, I will be more than
> happy to share, of course.
>
> Nico
>
> --
> Xorg exits when it receives an ACPI button/lid event
> https://launchpad.net/bugs/61746
>

Hi,

I tried the solution posted by Nico (http://librarian.launchpad.net/6385251/xserver-xorg-core_1.1.1-0ubuntu12.1%2Bacpifix_i386.deb) and unfortunately it didn't work for me. FYI, I am running edgy on a Dell Inspiron E1405 and have the latest BIOS available from the Dell support site.

Please let me know how I can fix this issue because whenever I close the lid, I have to resort to nothing but a hard reboot; and this is getting into my nerves :)

Thanks in advance.

This seems to be fixed in feisty now that it has xorg 7.2... I'm not getting any more issues since I updated to feisty.

marking fix released as per last comment. seems to still need a fix backported to edgy.

The xserver-xorg-core update from several hours ago once more ignores this bug and overrides once more the here released inofficial bugfix.

There IS a known FIX.

There is NO FIX RELEASED for edgy.

I hereby request an urgent backport of this fix to edgy.
(even if most, but not all will upgrade to feisty)

Please provide a proper SRU request according to https://wiki.ubuntu.com/StableReleaseUpdates, particularly with an explanation and origin of the patch (especially since it has not been confirmed to work in comment 50)

OK, as you (and the SRU process) wish.

Stage 1 - Propose
a. ubuntu-sru is subscribed
b. Problem is described here and in xorg upstream
 impact: xorg xrashes on acpi "video" event
c. fix released upstream
 https://bugs.freedesktop.org/show_bug.cgi?id=8776#c4
d. confirmed to work here by comments 32,33,36,38,40 and 42. (42 is THE ANSWER, so it has to be right).
    comment 50 provides not enough information and maybe describes an unrelated problem.

How to mark this bug as "Backport fix to releases"? What's next?

Hey y'all,

don't have the time & ressources to follow the above workflow right now, but as I do depend on my notebook computer, I rebuilt a patched package of the updated version. Thought I might as well share it with the rest of you...

For your information:
+ `dch' -i was used
+ debian/patches was used (never done it before)
+ it works for me
- due to lack of time, this package hasn't undergone intensive testing yet.

As always, feedback will be appreciated.

Luckily this is fixed in Feisty.
Still, it's awkward to see such an important bug not getting fixed for 2 (?) major releases of Ubuntu.

Let's hope our production machines won't need to go through such a nightmare ever again ;-)

All best,
Dan

Alright, approved for edgy-proposed.

No action here for a long time, and Edgy is becoming old.