Skip to contentSkip to conversation

programs vanish with 2.6.22+

DISCUSSION Markus 7 participants 17 messages Dec 6, 2007 – Dec 16, 2007

lore
Share & tools
lore

Markus

author
#0 lore raw

Hello!

Excuse my bad english, I am no native speaker. And please CC me, as I am not subscribed!

I am running an amd64 system with a stable gentoo. Today I upgraded to the kernel 2.6.22 and noticed many programs to disappear.
I use a kde desktop (3.5.7) and when amarok or kmail crash, normaly a "kde-crash-manager" pops up (but that happend very rarely), but not with 2.6.22+. It happens often and without any messagebox. kdesktop, xchat, kicker, konqueror crashed as well.
It happens more often under heavy load. But it can also happen when idle. Also I have the feeling that it happens more often on 2.6.23+ I might be wrong.

I tried 2.6.23 and the current 2.6.24-rc4, which all show this problem aswell!
What has changed in 2.6.21 -> 2.6.22 that could cause such things? Or who can tell that?
Can I help? How? (I use linux for years and can code, but I never debugged anything as big as the kernel...)

Markus

PS: Sorry thats so late, but I normaly stay on the distros stable kernel and only try vanilla when something is wrong ;)

author
#1 lore raw

Just forgot to mention:
dmesg gives nothing (/var/log/messages) the global X log aswell

and in the users .xsession-errors I have:
amarokapp: Fatal IO error: client killed
kdesktop: Fatal IO error: client killed
konqueror: Fatal IO error: client killed
The application 'xchat' lost its connection to the display :0.0;

I dont know what that means, but at least its something ;)

Markus

Quoted 3 lines

Hello!

Excuse my bad english, I am no native speaker. And please CC me, as I

am

Quoted 20 lines

not subscribed!

I am running an amd64 system with a stable gentoo. Today I upgraded to the kernel 2.6.22 and noticed many programs to disappear.
I use a kde desktop (3.5.7) and when amarok or kmail crash, normaly a "kde-crash-manager" pops up (but that happend very rarely), but not with 2.6.22+. It happens often and without any messagebox. kdesktop, xchat, kicker, konqueror crashed as well.
It happens more often under heavy load. But it can also happen when idle. Also I have the feeling that it happens more often on 2.6.23+ I might be wrong.

I tried 2.6.23 and the current 2.6.24-rc4, which all show this problem aswell!
What has changed in 2.6.21 -> 2.6.22 that could cause such things? Or who can tell that?
Can I help? How? (I use linux for years and can code, but I never debugged anything as big as the kernel...)

Quoted 3 lines

Markus

PS: Sorry thats so late, but I normaly stay on the distros stable

kernel

Quoted 2 lines

and only try vanilla when something is wrong ;)

#2 lore raw
Quoted 11 lines · Markus

Just forgot to mention:
dmesg gives nothing (/var/log/messages)
the global X log aswell

and in the users .xsession-errors I have:
amarokapp: Fatal IO error: client killed
kdesktop: Fatal IO error: client killed
konqueror: Fatal IO error: client killed
The application 'xchat' lost its connection to the display :0.0;

I dont know what that means, but at least its something ;)

Hm, no good idea, but some possible hints:

1. Ask Gentoo folks - if anyone has seen similar problems
2. RAM test - even if earlier kernels ran stable, the new one might stress your RAM differently
3. Are you sure nothing else has changed apart from the kernel?
4. You might try to recompile a vanilla 2.6.23+ kernel and enable as many kernel debugging options as you can...
5. Are you sure the installed kernel matches your CPU and your user-space 64- / 32-bit combination? Don't know whether a wrongly configured kernel could cause such problems though

Good luck
Guennadi

Quoted 35 lines

Markus

Quoted 3 lines

Hello!

Excuse my bad english, I am no native speaker. And please CC me, as I

am

Quoted 20 lines

not subscribed!

I am running an amd64 system with a stable gentoo. Today I upgraded to the kernel 2.6.22 and noticed many programs to disappear.
I use a kde desktop (3.5.7) and when amarok or kmail crash, normaly a "kde-crash-manager" pops up (but that happend very rarely), but not with 2.6.22+. It happens often and without any messagebox. kdesktop, xchat, kicker, konqueror crashed as well.
It happens more often under heavy load. But it can also happen when idle. Also I have the feeling that it happens more often on 2.6.23+ I might be wrong.

I tried 2.6.23 and the current 2.6.24-rc4, which all show this problem aswell!
What has changed in 2.6.21 -> 2.6.22 that could cause such things? Or who can tell that?
Can I help? How? (I use linux for years and can code, but I never debugged anything as big as the kernel...)

Quoted 3 lines

Markus

PS: Sorry thats so late, but I normaly stay on the distros stable

kernel

Quoted 2 lines

and only try vanilla when something is wrong ;)

Quoted 6 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

---
Guennadi Liakhovetski

author
#3 lore raw

Am Donnerstag, 6. Dezember 2007 schrieb Guennadi Liakhovetski:

Quoted 12 lines · Markus
Quoted 11 lines

Just forgot to mention:
dmesg gives nothing (/var/log/messages)
the global X log aswell

and in the users .xsession-errors I have:
amarokapp: Fatal IO error: client killed
kdesktop: Fatal IO error: client killed
konqueror: Fatal IO error: client killed
The application 'xchat' lost its connection to the display :0.0;

Quoted 5 lines
Quoted 1 line

I dont know what that means, but at least its something ;)

Hm, no good idea, but some possible hints:

1. Ask Gentoo folks - if anyone has seen similar problems

Noone gave a positiv answer so far.

Quoted 1 line

2. RAM test - even if earlier kernels ran stable, the new one might

stress

Quoted 1 line

your RAM differently

Will do that overnight.

Quoted 1 line

3. Are you sure nothing else has changed apart from the kernel?

Yes, I have some kernels installed and booted into them without changing anything else.

Quoted 1 line

4. You might try to recompile a vanilla 2.6.23+ kernel and enable as

many

Quoted 1 line

kernel debugging options as you can...

I try that next.

Quoted 1 line

5. Are you sure the installed kernel matches your CPU and your

user-space

Quoted 1 line

64- / 32-bit combination? Don't know whether a wrongly configured

kernel

Quoted 1 line

could cause such problems though

Everything is "safe". Its a native 64-bit system with "safe" flags.

Quoted 4 lines

Good luck
Guennadi

I will write when I did a ram-test and am running a 24-rc4 with full-debug.
Will also try things from the gentoo-users...

Is there any userspace software that is needed with newer kernels?

Markus

Quoted 6 lines
Quoted 6 lines

Markus

Quoted 3 lines

Hello!

Excuse my bad english, I am no native speaker. And please CC me,

as I

Quoted 4 lines
Quoted 4 lines

am

Quoted 3 lines

not subscribed!

I am running an amd64 system with a stable gentoo. Today I

upgraded to

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

the kernel 2.6.22 and noticed many programs to disappear. I use a kde desktop (3.5.7) and when amarok or kmail crash,

normaly

Quoted 1 line
Quoted 1 line
Quoted 1 line

a "kde-crash-manager" pops up (but that happend very rarely), but

not

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

with 2.6.22+. It happens often and without any messagebox. kdesktop, xchat, kicker, konqueror crashed as well. It happens more often under heavy load. But it can also happen

when

Quoted 1 line
Quoted 1 line
Quoted 1 line

idle. Also I have the feeling that it happens more often on

2.6.23+ I

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

might be wrong.

I tried 2.6.23 and the current 2.6.24-rc4, which all show this

problem

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

aswell!
What has changed in 2.6.21 -> 2.6.22 that could cause such things? Or who can tell that?
Can I help? How? (I use linux for years and can code, but I never debugged anything as big as the kernel...)

Markus

Quoted 8 lines
Quoted 8 lines
Quoted 1 line

PS: Sorry thats so late, but I normaly stay on the distros stable

kernel

Quoted 2 lines

and only try vanilla when something is wrong ;)

--
To unsubscribe from this list: send the line "unsubscribe

linux-kernel" in

Quoted 8 lines
Quoted 4 lines

the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

---
Guennadi Liakhovetski

author
#4 lore raw
In reply to Guennadi Liakhovetski

Hi again!

The memtest ran 14 passes (~10h) without an error.

I now have a 2.6.24-rc4 with some debug-options turned on, waiting for something to happen... can I just leave it untill a window disappears or do I need to manually enable something or run some user-space app?!

Markus

#5 lore raw
Quoted 7 lines · Markus

Hi again!

The memtest ran 14 passes (~10h) without an error.

I now have a 2.6.24-rc4 with some debug-options turned on, waiting for something to happen... can I just leave it untill a window disappears or do I need to manually enable something or run some user-space app?!

It depends - different options have it differently. Most simple ones are just compile-time, so, you don't have to enable them. Look in "help" for respective debug-options.

Thanks
Guennadi
---
Guennadi Liakhovetski

author
#6 lore raw

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice.
Or a hit, in what direction I should look. Because its really nasty to not being able to use a current kernel.

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it to disappear.

Just talk to me, I am not able to do this on my own...

Markus

Quoted 7 lines · Markus
Quoted 6 lines

Hi again!

The memtest ran 14 passes (~10h) without an error.

I now have a 2.6.24-rc4 with some debug-options turned on, waiting

for

Quoted 1 line
Quoted 1 line

something to happen... can I just leave it untill a window

disappears

Quoted 1 line
Quoted 1 line

or do I need to manually enable something or run some user-space

app?!

Quoted 2 lines

It depends - different options have it differently. Most simple ones

are

Quoted 1 line

just compile-time, so, you don't have to enable them. Look in "help"

for

Quoted 7 lines

respective debug-options.

Thanks
Guennadi
---
Guennadi Liakhovetski

#7 lore raw

fre, 07 12 2007 kl. 23:52 +0100, skrev Markus:

Quoted 6 lines

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice.
Or a hit, in what direction I should look. Because its really nasty to not being able to use a current kernel.

Quoted 8 lines

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it to disappear.

Just talk to me, I am not able to do this on my own...

If you feel like you are able to tell whether a specific kernel version is buggy or not, you might want to try to bisect it. See Documentation/BUG-HUNTING in the sources, and please ask.

Simon Holm Thøgersen

author
#8 lore raw

I try that, but it will take a lot of time!

Markus

Quoted 5 lines

fre, 07 12 2007 kl. 23:52 +0100, skrev Markus:

Quoted 4 lines

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice.
Or a hit, in what direction I should look. Because its really nasty

to

Quoted 6 lines
Quoted 6 lines

not being able to use a current kernel.

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it

to

Quoted 5 lines
Quoted 4 lines

disappear.

Just talk to me, I am not able to do this on my own...

If you feel like you are able to tell whether a specific kernel

version

Quoted 7 lines

is buggy or not, you might want to try to bisect it. See Documentation/BUG-HUNTING in the sources, and please ask.

Simon Holm Thøgersen

#9 lore raw
Quoted 3 lines · On Sat, 8 Dec 2007 13:12:14 +0100 Markus <[email protected]> wrote:

I try that, but it will take a lot of time!

Markus

This problem remembers me something...

http://groups.google.com/group/linux.kernel/browse_thread/thread/85859519ffec7dc8/591a0b3a05bd3596?lnk=gst&q=konqueror+vanish#591a0b3a05bd3596

Are you the same Markus? (or it's just a coincidence?)

It seems the same BUG...

Signature · 2 lines
	Paolo Ornati
	Linux 2.6.24-rc4-g7e1fb765 on x86_64
author
#10 lore raw

Well, no I am not the same markus. And I found that before, but I thought it was something about cfs and imo that made it into linus-tree in .23 not .22.
But I should perhaps try to change my name, perhaps that fixes it -.-

Markus

PS: am currently doing a bisect, thats really bad: third bisect and the kernel is unusable.

Quoted 10 lines
Quoted 3 lines · On Sat, 8 Dec 2007 13:12:14 +0100 Markus <[email protected]> wrote:

I try that, but it will take a lot of time!

Markus

This problem remembers me something...

http://groups.google.com/group/linux.kernel/browse_thread/thread/85859519ffec7dc8/591a0b3a05bd3596?lnk=gst&q=konqueror+vanish#591a0b3a05bd3596

Quoted 5 lines

Are you the same Markus? (or it's just a coincidence?)

It seems the same BUG...

author
#11 lore raw
In reply to Paolo Ornati

Well, after that I tried to verify that 22 is really fine. Unfortunatelly under heavy load its the same... even with .19 (the oldes kernel I have still installed)
I just noticed it, because its harder to produce it on .22 and before. You guys tuned the kernel so it was easier.
So it must be something in userspace, please excuse me ;)

Markus

#12 lore raw
In reply to Markus
Quoted 11 lines · Markus

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice.
Or a hit, in what direction I should look. Because its really nasty to not being able to use a current kernel.

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it to disappear.

Well, you could attach strace to all likely crash candidates like

strace -etrace=none -o/tmp/<pid>.trace -p<pid>

which would at least tell you what signal it caught...

good luck
Guennadi

Quoted 15 lines

Just talk to me, I am not able to do this on my own...

Markus

Quoted 7 lines · Markus
Quoted 6 lines

Hi again!

The memtest ran 14 passes (~10h) without an error.

I now have a 2.6.24-rc4 with some debug-options turned on, waiting

for

Quoted 1 line
Quoted 1 line

something to happen... can I just leave it untill a window

Quoted 17 lines

disappears

Quoted 1 line
Quoted 1 line

or do I need to manually enable something or run some user-space

app?!

Quoted 2 lines

It depends - different options have it differently. Most simple ones

are

Quoted 1 line

just compile-time, so, you don't have to enable them. Look in "help"

for

Quoted 7 lines

respective debug-options.

Thanks
Guennadi
---
Guennadi Liakhovetski

---
Guennadi Liakhovetski

author
#13 lore raw

Well, just tried it. Started a dozen konquerors and attached strace to everyone. When one disapeared, I only got a "Process 9246 detached", nothing else is printed or written in the log.

Markus

Quoted 6 lines · Markus
Quoted 5 lines

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice.
Or a hit, in what direction I should look. Because its really nasty

to

Quoted 6 lines
Quoted 6 lines

not being able to use a current kernel.

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it

to

Quoted 24 lines
Quoted 1 line

disappear.

Well, you could attach strace to all likely crash candidates like

strace -etrace=none -o/tmp/<pid>.trace -p<pid>

which would at least tell you what signal it caught...

good luck
Guennadi

Quoted 13 lines

Just talk to me, I am not able to do this on my own...

Markus

Quoted 7 lines · Markus
Quoted 6 lines

Hi again!

The memtest ran 14 passes (~10h) without an error.

I now have a 2.6.24-rc4 with some debug-options turned on,

waiting

Quoted 7 lines
Quoted 7 lines

for

Quoted 1 line
Quoted 1 line

something to happen... can I just leave it untill a window

disappears

Quoted 1 line
Quoted 1 line

or do I need to manually enable something or run some user-space

app?!

Quoted 2 lines

It depends - different options have it differently. Most simple

ones

Quoted 2 lines
Quoted 2 lines

are

Quoted 1 line

just compile-time, so, you don't have to enable them. Look

in "help"

Quoted 14 lines
Quoted 10 lines

for

Quoted 7 lines

respective debug-options.

Thanks
Guennadi
---
Guennadi Liakhovetski

---
Guennadi Liakhovetski

#14 lore raw
Quoted 5 lines · Markus

Well, just tried it. Started a dozen konquerors and attached strace to everyone. When one disapeared, I only got a "Process 9246 detached", nothing else is printed or written in the log.

Markus

Hallo Markus

Whenever the connection to the X server is dropped the application using that connection will be terminated by the client library error handler (KDE/Qt/X11).

Probably the applications simply drop out of their event loop and terminate, you should see some X11 related error messages in

.xsession-errors

My guess would be that some of the eye-candy in KDE is using DRI / OpenGL features that might be the cause for mailfunction.

I would suggest you start by disabling DRI / AIGLX rendering in your xorg.conf and see if your applications work without all those nifty features.

Below you can find some settings to disable AIGLX, DRI and composite rendering.

They should be integrated in the configuration sections I quoted. This is not a complete configuration file.

Best regards,
Patrick

--- <snip> ---

Section "ServerFlags"
Option "AIGLX" "false"
EndSection

Section "Module"
Disable "dri"
EndSection

Section "Extensions"
Option "Composite" "false"
EndSection

--- <snip> ---

#15 lore raw
In reply to Markus
Quoted 3 lines · Markus

Well, just tried it. Started a dozen konquerors and attached strace to everyone. When one disapeared, I only got a "Process 9246 detached", nothing else is printed or written in the log.

You could try an ltrace instead, and see if it's one of the linked libraries having a problem.

#16 lore raw
In reply to Markus
Quoted 9 lines · Markus

Well, now some windows vanished, but no additional messages were produced by kernel. When somebody could tell what I exactly need to do... would be nice. Or a hit, in what direction I should look. Because its really nasty to not being able to use a current kernel.

I already rebuild the whole system, as suggested by the gentoo-devs, without success.

I could also try to debug/strace/whatever the apps and wait for it to disappear.

Just talk to me, I am not able to do this on my own...

Try running under GNOME if you can, this sounds as if it may be an X problem.

Signature · 3 lines
Bill Davidsen <[email protected]>
  "We have more to fear from the bungling of the incompetent than from
the machinations of the wicked."  - from Slashdot