Kubuntu freezes when desktop effects are enabled

Could you please attach a copy of the output of lspci -vvnn (two v's, not a w) and a copy of your /var/log/Xorg.0.log please? Thanks.

Note that the file /var/log/Xorg.0.log was created by the current X session which has KDE desktop effects disabled (and isn‘t crashed, of course). If you need a /var/log/Xorg.0.log from a session with desktop effects enabled or even from after a freeze, please tell me.

Hi there,
looks similar to Bug #350766 which I have opened. I just marked mine as a duplicate of yours.
greetz

Hi, lspci..here we go..

It looks this issue doesn't exist on my RV350 (Radeon 9600). I've upgraded Kubuntu 8.10 to 9.04 (x86) few days ago and since then I've been running my computer for few hours a day with KDE desktop effects just fine. I tried some OpenGL games and apps running for few hours, too. Since then I haven't noticed any problems with "radeon" driver.
Might it be an issue for R4xx and R5xx cards?

This suggest it's not a kwin specific problem like suggested in Bug # 350766 but a driver related one.

I have freezes when kwin desktop effects are enabled, but I am using nvidia. The last one came even before I could login. So, if it's a driver issue, both Ati an nvidia seem bugged.

Those random (when desktop effects are on) freezes occur also on my machine (Jaunty/KDE4.2/KWin) which has Intel i915 chipset integrated graphics. So it does not seem to be GPU vendor-related.

I've tested kwin now. The desktop effects use opengl (at least in my case) and I get the same freezing as described here and in Bug # 350933. I'm still not sure if this problem is kwin specific or opengl specific.

Interestingly, the gnome desktop effects work fine. What renderer is used by gnome?

I'm pretty convinced now that Bug # 350933 and this one result from the same error. Do you think it should be marked as a duplicate?

@ Thomas

I think that this Bu report, #350933 and #350766 (the one I have reported) have a lot in commom - so yes I think
thay look like duplicates...

The status of this bug is currently “incomplete”. What does this mean? That there is incomplete information about this bug? This would no longer be true since I provided the requested information 4 days ago. So if “incomplete” really means “incomplete info” then please set the status to something different than “incomplete” in order for this bug to be worked on.

This is off-topic: On my duplicate, the "incomplete"-label was also not removed after I provided additional information. But it should be considered that
* At least in my country (Germany) there are easter holidays, so not every (of the paid) programmers might be at work
* Jaunty will be released soon. There are certainly worse problems than issues with desktop effects witch need to be solved until then.
* Everyone of us is welcome to work at this bug - we are the community. Everyone of us (I don't know if the reporter himself, but the rest) can change the bug status to "confirmed". The problem at my side is, I don't know what the "standard" requirements are to fulfill "confirmed" so I won't touch this.

If have more questions like this: They certainly are important, but they should be placed into Launchpad's "Answers".

Desktop effects are one of the outstanding features of KDE 4. For me, they are one of the main reasons for switching to Kubuntu. And switching them on causes complete system crashs at unpredictable times. So in my opinion, bugs like this one should be fixed in a stable release.

To put it together:
1. I saw the same freezing issue also on the stable OpenSuse 11.1 with KDE 4.2 installed separately (as 11.1 has only 4.1) from the development (factory) branch.
2. So far, it happens on ATI, Intel and nVidia video hardware.
3. Crashes are unpredictable. Differences may be extreme. One day everything was great, another day - it froze almost immediately after launch.
4. It takes place on both: KDE 4.2.1 and KDE 4.2.2.
5. Desktop effects are a great feature.

I have this very same issue now on my machine as well. I had upgraded from 8.04 to 8.10 to 9.04. As a result of encountering very similar issues mentioned in https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/350120, I decided to do a clean install of 9.04 by wiping out all my partitions except my /home partition. I was hoping that would address my problem, but instead I get complete random freeze ups described in this report. In hopes of getting this issue resolved I will post the information from apport-collect soon.

Is there any progress on this bug?

Ok, i try compiz-fusion window manager in KDE and no freeze. The problem is Kwin.

Hi
I have been using Compiz desktop effects for 48 hours also no problems with freezes.
System runs just fine.

I am experiencing the same random hangs when using kwin with desktop effects: mouse movement possible, no updates or response of screen, "clean" shutdown only possible using ACPI (power button).

I am running a fully up to date Jaunty using radeon with EXA acceleration. Personally I find that the crashes do not occur (or maybe less frequently) when using XAA instead of EXA. I've been experiencing this bug for a long time (since KDE 4.1 was available for Kubuntu). I have frequently tried using svn builds of the ati driver (from xorg-edgers and tormodvolden ppa) but they always had the same problem.

Other than the random freezes, radeon + kwin effects works great, so it'd be really nice if that last issue disappeared.

I have attached my lspci output.

I did some research and come to this conclusion:
On Kubuntu Alpha 6 everything worked fine, but with beta came kernel update and that is the time when problem has start to show up.
I tried to use Compiz and worked flawlessly.
Also when using Compositing type: XRender don't have any problems with randomly freezing.
Hope this help to someone who is smarter then me, and solve the problem.

@Bojan:

I just tried XRender, in my system I get massive graphic distortion on my windows when activated, though it
is impossible to test it any further....
I attached a screenshot how windows look when using XRender, the distorted Windows in the front middle ist actually "System Settings" window :-)

@yuhu
I also have some graphic distortion with XRender, but not that scary.

Hi,
Since yesterday I am using the packages provided by Tormod Voldens PPA which include at driver 6.12.2 adn I think updated Mesa driver as well.
And up to now I didnt have a single freez using Kwin compositing effects, some one wanna give it a try?
greetz

This don't help me. :-(

Did you add the PPA to your apt sources list or did you just
download the xorg-ati packge?
Because when you use the PPA via apt it updates a couple of packages, so I cannot know which
one actually fixed it .... at least for the time beeing, maybe I will get a freeze later on - who knows...

I already have this in my sources list:
deb http://ppa.launchpad.net/tormodvolden/ppa/ubuntu jaunty main

Do i also need this:
deb-src http://ppa.launchpad.net/tormodvolden/ppa/ubuntu jaunty main

I am also experiencing this with a Radeon Pro 9600 (RV350).

Forgot to add, I can still access my machine using ssh, but even after restarting X the keyboard is non responsive and the screen is blank. Only sysreq works.

I've said before this issue doesn't seem to exist on my hardware but after a bit more testing I found it does exist, too.
But if I use Kubuntu and desktop effects only for konsole/yakuake, konqueror, kontact (kmail), plasma applets etc. kde4 apps is stable for me and I haven't experienced freeze (yet). I use my computer for few hours a day. But I can reproduce hangs by for example use k3b 1.0.5 (kde3 app) and for example tvtime for analog tv watching using xv at the same time (kdetv doesn't work in kubuntu 9.04 (regression in 9.04 after upgrade from 8.10 - described in bug #325874) but it occurs after some time and using dialogs in k3b. I've found that it occurs (I think) with gtk apps too on kubuntu (it happened to me just one time or maybe twice). It might be because of one of desktop effects plugins (one of the enabled by default). After hang I can still use sysrq (mostly).
Graphic card: Sapphire Radeon 9600 256MB AGP 8x (RV350).

I can confirm this behaviour, system locks up or screen goes completely dark with a hard reset as the only option left.
This is on a Laptop with an Ati Radeon 9600 Mobility M10 (RV350) card using the radeon driver and kwin's compositing effects enabled.

The lock-ups occur when opening random applications, so far it happened to me while opening kwallet, login-manager, kate, kpackagekit and others.

Also, X sometimes uses up to 50% of the cpu-time.

Hi

In Fedora forum they recommended / found a fix that involved acpi functions of the kernel.
In addition most people reported that this issue is doesn t occour with older kernel i.e. 2.6.27-11.
There seems to be a fixed xorg driver for intel cards (last page in the following link).
I by myself can only confirm that newer kernel version .i.e 2.6.30 RC2 do still have the bug.....
Anyone wit mor eknowledge might want to read the thread and see what the guys have found out - sadly I didnt get it all :-(

http://forums.fedoraforum.org/showthread.php?t=205950&page=7

With this kernel everything works fine:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28.9/

I am experiencing the same issue with my card, an ATI Technologies Inc RV370 [Sapphire X550 Silent]: black screen after opening a new window, like a skype chat or the network manager config. Touching wood, it appears the workaround of disabling the window effects is... effective.

Have you try to install some other kernel? By other i mean new-one.
Also you can try to boot your system with:
 noapic acpi=off
parameters.

What do you mean with a “new” kernel? Which version? From the jaunty APT source or from some special source? If the latter, how about security fixes?

Does the ACPI/APIC thing have something to do with this bug at all? This other bug report where disabling ACPI/APIC was suggested was about freezes which were not necessarily related to X11 and desktop effects, AFAIK.

Like i said with this kernel everything working fine:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28.9/

I didn't try to insert noapic or acpi=off because i have old motherboard.

I've the same problem. I use a Toshiba M50-148 laptop with a Radeon X600 Mobile (RV280).
I have attached my lspci -vv output.

I've the same problem. I use a Toshiba M50-148 laptop with a Radeon X600 Mobile (RV380).
I have attached my lspci -vv output.

I have installed the 2.6.28.9 kernel and the KDE4 desktop effects still hang on my laptop :/

I'm sorry to heard that. You can try some older kernel, there is no real solution for this.

Try the kernel 2.6.29.1 it worked for me (but the problem is back with th 2.6.30...)

Thanks for the hints. I will try various options and eventually post back whether they brought any success.

Kernel 2.6.29.1 didn’t work for me. Same problems as with the Ubuntu kernel. Maybe it actually doesn’t even work if you think it does since these crashes seem to occur rather unpredictably. Sometimes they occur very frequently, sometimes there are no crashes for one or two days.

Hello,

is there any progress in fixing this bug?

Best wishes,
Wolfgang

On KDE dev I have found that someone compiled xorg 1.5.2 packages for jaunty which fixed the issue.

Here is the story:

http://www.kdedevelopers.org/node/3942

You are free to try :-)

http://www.davidfaure.fr/2009/xorg_debs.tar.bz2

I for myself havent tried it yet because with kernel 2.6.29 I havent had any freeze for 6 whole days...

Note that dfaure’s problem is not necessarily the same as the problem(s) described here. dfaure wrote about red dots which I never experienced. On the other hand, he says that he didn’t experienced problems with NVidia while others reported here that they did.

Anyway, I’ll probably try X.org 1.5.2 and see if this solves the problem for me.

Would it be a good idea if the Ubuntu developers make official X.org 1.5.2 packages (including ati and fglrx drivers) which are available via some APT repository and also get security fixes?

@ Wolfgang Jeltsch:

I know that dfaure's problem isn t necessarily the same problem as ours. But there are quite a couple of Kwin / Xorg freeze problems when using compositing and / or opengl programs (see duplicates of this bug, bug #348450 and probably more...).
The problem itsself seems to be difficult to debug, some people are fine using Kernel 2.6.29 but others aren't, so the "freeze bug" in general might or might not have something in common, with different effects on different machines.

btw:
Tormod has new radeon drivers in his PPA, you might wanna try them.
They didnt work for me I got a freeze after a couple of hours :-(

I have Radeon Mobility x1600 and had the same problems with freezing.

After installing kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.29.2/ and all freezes with compositing "on" are gone. It is working.

I have been using 2.6.29.1 for a couple of days and the general impression is: better though far from good. The freezes just seem to be much less frequent (they still happen on my machine). It might be possible to work several days in a row without a single freeze with this version.

As this phenomenon is not ATI-specific (I have an Intel i915, for instance), I am skeptical whether it could be a driver issue. Still, it may be possible that these are just various issues having common symptom - a freeze. Well, perhaps it's not a bug but a feature - a sort of Win98 First Edition stability mode (they might call it "Win9x Effects") :-]

2.6.29.2 is next in the line.

jpkx1984, you can try 2.6.28-12.43 as jtholmes suggests here:
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/348332/comments/35
For me 2.6.29.2 is working. It has been more than 1 day and no freeze. If I have one I will post a message here.

This seems related, a into looks like it fixed the freezing problems
https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/368049

I don’t understand this. The page you referenced should be a bug report (of bug #368049) but it actually describes the fix of a bug. Why is a fix reported as a bug? Where are the corresponding bug reports?

And when was this fix commited? There was a mesa update a couple of days ago which was said to fix X server crashes. However, it removed a patch that was added at the beginning of April while I experienced crashes already at the end of March. Is this “bug” #368049 this mesa fix? If yes then it won’t help me. And it won’t help all the others who still experienced problems after this fix has gone into Jaunty a couple of days ago.

@ Wolfgang:

Fix was committed for jaunty proposed on 05/05/09 in mesa versions 7.4-0ubuntu4 it it seems to be for r300 driver only.
So now u can check if you already have those packages.

If this fix does not resolve our problem we have to keep working to figure it out...

tormod drivers:
xserver-xorg-video-radeon_6.12.2.99+git20090504.da58e351-0ubuntu0tormod_amd64
xserver-xorg-video-radeon_6.12.2.99+git20090506.0c3541f7-0ubuntu0tormod_amd64

Do not resolve the freeze issue, they merely make it appear in kernel 12.6.29.2 which is "freeze free" with standard jaunty packages for me.

cya

I checked the mesa changelogs again and discovered that the fix that I had in mind (the one that fixed the problem introduced at the beginning of April but didn’t help me) appeared in mesa 7.4-0ubuntu3. This is also the mesa version that is currently installed on my machine.

However, my chip seems to be an RS480. So mesa 7.4-0ubuntu4 won’t help me, would it?

Now I have googled for “r300 rs480” and discovered that both things seem to be related. RS480 is the name of the chip. What is R300? Is it not a chip name? What is it then? Is it true that RS480 somehow belongs to R300?

Since I have an RS480 and therefore seem to have something R300-related, I installed everything new from jaunty-proposed. However, this doesn’t include mesa 7.4-0ubuntu4 but only mesa 7.4-0ubuntu3.1. Where is 7.4-0ubuntu4?

@Wolfgang Jeltsch
RS480 is the technical chip name knowed as "Radeon Xpress 200 IGP" for the user.
R300 is the third generation of Radeon graphics chips from ATI Technologies.
You can immagine that R300 is a common architeure of all R300 chips so called.
I take information from here:
http://en.wikipedia.org/wiki/Radeon_R300
http://en.wikipedia.org/wiki/Comparison_of_ATI_Graphics_Processing_Units#Radeon_R300_PCIe_.28X3xx.2C_X5xx.2C_X6xx.29_series

Sorry I am mistaken the bug was fixed in the package mesa - 7.4-0ubuntu4 for KARMIC!

For jaunty it is fixed in mesa 7.4-0ubuntu3.1 as you can see here (it is the fixed talked about in #368049):

https://launchpad.net/ubuntu/jaunty/+source/mesa/7.4-0ubuntu3.1

@3ntix:

Thank you for that information. Actually, my chip seems to be an RS482. It’s an Radeon Mobility XPRESS 200M (PCIE) or something like that.

So I’m using a fixed version now, and what shall I say? I just had a crash again!

However, this kind of crash was a bit different from what I originally reported. The screen went black(?) for about 0.1 seconds and then the normal picture reappeared. And the system hung. I think, this time, not even the mouse pointer responded anymore.

I really would like to know how much time it will probably take to fix these crash problems. Are there any estimations?

I’m currently thinking about installing Debian lenny with KDE 4.2.2 backports. This would give me version 1.4.2 of the X.org X server so that I probably won’t experience these crashes. In addition, I would be able to use fglrx with my “old” card which might give me smooth desktop effects. With the open source radeon driver, desktop effects are usually jerky, and version 1.6 of the X server only supports fglrx versions that don’t support R300 anymore.

I’d really be interested to hear why Jaunty had to come with version 1.6 of the X server. Kubuntu Jaunty currently reminds me of Windows: Don’t install until the first service pack is out. I’m really astonished that something that’s crashing and crashing again is considered a stable version and officially released.

I wanted an easy-to-install Linux with KDE 4.2. I thought that Kubuntu was the way to go. But what I have experienced during the last weeks is really annoying. I didn’t want to do some complicated installing – Debian plus backports plus proprietary driver – but at the moment it seems like this would be the best solution. :-(

So please take this as a plea for a Kubuntu Karmic which is more stable and not so bleeding-edge.

Have you ever get a look of this bug report?
https://bugs.launchpad.net/ubuntu/+bug/368461
I've opened it. I think that there are more than one thing that cause
a X crash. I've seen tree type of different crash (maybe four).
As I write in #368461:
(1) system freeze, the screen propose the last seeing, the mouse can
move, pc don't respond with keyboad, ssh on (and safely reboot). This
behaviour is related to this bug
(2) the screen become black and the pc don't respond at all: no mouse,
no keyboard, no ssh (hard reboot). This is related to #368461
(3) just a crash of kdm. Then I can login into kdm-login-interface and
the pc is up.
I've try to catch a backtrace when system crash by installing
libgl1-mesa-dri-dbg libc6-dbg xserver-xorg-core-dbg
xserver-xorg-video-ati-dbg packages but nothing.

Have you ever get a look of this bug report?
https://bugs.launchpad.net/ubuntu/+bug/368461
I've opened it. I think that there are more than one thing that cause
a X crash. I've seen tree type of different crash (maybe four).
As I write in #368461:
(1) system freeze, the screen propose the last seeing, the mouse can
move, pc don't respond with keyboad, ssh on (and safely reboot). This
behaviour is related to this bug
(2) the screen become black and the pc don't respond at all: no mouse,
no keyboard, no ssh (hard reboot). This is related to #368461
(3) just a crash of kdm. Then I can login into kdm-login-interface and
the pc is up.
I've try to catch a backtrace when system crash by installing
libgl1-mesa-dri-dbg libc6-dbg xserver-xorg-core-dbg
xserver-xorg-video-ati-dbg packages but nothing.

Maybe it’s related to this other bug. However, as I’ve said, my screen does not turn to black permanently. It’s just a short flicker. The current image disappears and reappears immediately afterwards.

@3ntix
As I have written above on 2009-05-05 I do agree with you.

@ Wolfgang:
Well Ubuntu and Fedora are a little "on the edge" when it come to new packages and kernels this is usually something good but can also lead to some problems as wee see here.

Considering that there have been bug reports of these freeze issues well before the date of release (your report and the one I submitted) I do agree that there could have been more progress in solving the bug.

Btw.:
Despite the first post following the bug report I havent seen any Xorg maintainer working on this issue - please proof me wrong if I am.....

The new mesa packages ubuntu 3.1 for jaunty seem to introduce some sort of input lag on my system when compositing is activated.

It can be reproduced with the following:

open kpackage
search for "kde"
move with the mouse ober the results

Usually the option underneath the mouse gets highlighted pretty much immediately.
With the mesa packages it takes over a second and sometimes more.

Furthermore it can be seen while browsing with firefox, the mous needs some time to change from mouse pointer to "hyperlink click"..very annoying delay....

If someone can confirm this behaviour I would report it as a bug to the correspnding mesa release.

@juhuu, can you please attach the output of "tail --lines=5000 /var/log/dpkg.log"?

Also, can you uninstall only the 3.1 update and see if that issue goes away again? If you havn't cleaned out your apt cache the corresponding old DEBs should still be on your machines so you can uninstall using:
sudo dpkg -i /var/cache/apt/archives/*mesa*0ubuntu3_*.deb

Then logout and login again and see if the issue is gone?

If it goes away, can you then install the patch again (and double check) using:
sudo dpkg -i /var/cache/apt/archives/*mesa*0ubuntu3.1_*.deb
(not entirely sure but that _should_ work I guess)

@ Martin

I have already gone back to the standard packages. The problem has gone away as far as I can tell, though I now know what to look for I can still see it a little - but that might have been like that all the time :-)

When I get back home I reinstall the 3.1 packages again, to see if the issue re-appears.

cya

Same issue there. The system is very unstable when using KWin desktop effects.

My system :
Pentium M 1.7 GHz
Radeon 9700 Mobile

Kernel :
Linux aubin 2.6.28-11-generic #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009 i686 GNU/Linux

@ Martin

Hi
I have installed the 3.1 mesa packages yesterday. Until this morning everything was ok so far.
But just recently the system input got sluggish again.
The reason seems to have been the python process running at 50% CPU usage and 500MB memory, after I terminated the process system was fine again
Sorry for the trouble, I just didnt notice that before!?

Hi, I'm currently trying to understand the impact (and number of affected users) of a certain -ati lockup bug that some of you may or may not be affected by.

@Julien Aubin, @Heinrich Janzing, @Jose Bernardo, @Thomas Lauckner,
@Bojan Kostić, @Tomasz Czapiewski, @Wolfgang Jeltsch, @juhuu, @lanzen,
@jpkx1984, @jpkx1984, @DA, @Gobnuts, @aboaboit, @3ntix, @humufr, @aya, @Lafa,

It would be extremely useful if the people listed above booted the ubuntu jaunty live CD (not kubuntu) on their computers with ATI graphics cards, and then tried opening the website http://mundoplus.tv/ to see if that locks up your machines.
Booting from the live CD is nice because it means we can be absolutely sure that you are all using the exact same software for testing and that no configuration files or remains from upgraded intrepid installtions taint the results of this test.

If your machine does does indeed lockup when you visit http://mundoplus.tv/ using Firefox, then please post a confirmation as a comment and when confirming please also include the output of "lspci -nn | grep VGA".

Thank you.

Confirmed. I had a lockup using jaunty ->without<- live cd.
$ lspci -nn | grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon Mobility X1400 [1002:7145]
$ cat /etc/X11/xorg.conf|grep Driver
 Driver "ati"

Confirming lockup ->with<- live cd (downloaded and burnt today)
Mouse remains movable, everything else is locked up.

$ lspci -nn | grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon Mobility X1400 [1002:7145]

I did not have live CD around, so I just clicked the url (firefox) and the system locked. The pointer was moving jerkily, but I could ssh to the locked box from another machine to reboot it.

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc M56P [Radeon Mobility X1600] [1002:71c5]

Hi

Just tried it with Live CD and it locks up shortly after the first graphics of the page were loaded.
mouse can still be moved but system had to be hard-reset.

Status: confirmed with Live CD and without

Here is the output:

ubuntu@ubuntu:~$ lspci -nn | grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon Mobility X1400 [1002:7145]

After the bootup with the live CD, the Desktop ran without compositing enabled. The website ran fine, although in Konqueror (couldn't install firefox in the live CD environment).
As soon as I enabled the compositing effects, the screen froze with the mouse moving jerkily.

Status: confirmed with Live CD

$ lspci -nn| grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10] [1002:4e50]

$ cat /etc/X11/xorg.conf|grep Driver
produced no output, so I assume it used the ati driver

Ooooops, sorry. The whole procedure was done with the Kubuntu live CD. I'll grab the ubuntu live CD and try it again with ubuntu. Results will follow.

I have tried, accessing http://mundoplus.tv/ from the Ubuntu Live CD. The system hangs before the contents of the browser window have changed. The progress bar on the bottom right stays at about 80 %. The mouse pointer moves only jumpy and is not animated anymore (i.e., the animation halts). The only way, I can shut down the system is pressing Alt + SysReq + REISUB.

The output of lspci -nn | grep VGA is this:

01:05.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon XPRESS 200M 5955 (PCIE) [1002:5955]

I confirm the same here with "http://mundoplus.tv/" in Firefox - freeze type (1).

$ lspci -nn | grep VGA
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 AP [Radeon 9600] [1002:4150]

I can reproduce the behavior with and without the livecd (ubuntu in both cases, NOT KUbuntu). The system hang while the mouse still responses to my input. On my current install Compiz and Metacity composite are OFF but with mondoplus website the system still hang. With Compiz ON the system hang often and there's no way to go anywhere (even TTY) and mouse cursor is not working. The model of my ATI is:

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10] [1002:4e50]

If you need more test or more behavior to test just ask, I'm glad to help to resolve this annoying problem.

Fabio

After checking carefully, I believe this bug report (and also many of the people who commented on this bug) is a duplicate of bug 371279 (which is already reported upstream).

No more information is required now except one thing: information about the exact version when the bug was introduced. If you know how to do upstream git bisection this could definitely help us get closer to a solution.

Once/if this bug is fixed upstream, I will prepare an update for Ubuntu (first karmic and possibly also jaunty depending on how big/disruptive changes are required to fix the bug).

I have the same problem WITH ubuntu 9.04 LIVE and WITH my kubuntu (with proposed update and kde 4.2.3).
Here is the lspci
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc M24 [Radeon Mobility X600] [1002:5462]
Why this report is related to #371289? I think that this is a different bug.

@ Martin

Hi,
I am glad that we are getting somewhere after walking in darkness for almost 6 weeks.
After I have read bug report #371279 and upstream bug report #21598 I doubt that this bug is an exact duplicate.
The main reason why I believe this is that in 371279 and in upstream option XAA can be used to avoid the freezes.

description from 371279:
REPRO STEPS:
  1. (using firefox) visit website http://mundoplus.tv/

- Note: This bug goes away completely when you set "DRI" "off".
+ Note: This bug goes away completely when you set "DRI" "off" or if you
+ use "AccelMethod" "XAA".

However Wolfgang has reported that using XXA will also result in crash (black screen) and I can confirm that on my system.
Therfore I am not very sure if this really is a duplicate?

But well I hope we will figure it out :-)

..and please fix it in Jaunty too, otherwise I will probably go back to windows a third time...and I really dislike that idea...

@juhuu, yea that's strange actually. For now, let me updated the main bug to clarify this.

It might of course be that we're dealing with two different bugs being triggered by the same repro steps (one XAA bug that causes a black screen on Wolfgang's hardware and another bug that causes a lockup for EXA). It's extremely hard to know for sure. My gut feeling (and it's not much more than that) tells me that once we figure out what's special about that webpage we can fix both bugs. I've been trying various things here so far, like for example using wget to download all images off that site to see if there any of them can cause the crash in isolation but no luck so far. I've also tried various versions of the upstream DDX (I was hoping we'd find a range to bisect) but unfortunately I think that the default operational mode for the -ati driver was XAA in intrepid so it might very well be that EXA has been freeze free in the past (and in that case we can't bisect and that makes it a lot harder). I'm not sure more data and/or bug sorting can help us get closer to a fix right now. Personally, I'll focus on testing different versions of the driver now (6.12 branch master first and then maybe the radeon rewrite branch if I can get that to build). Finally, whether we can ship an update for jaunty depends on how disruptive / risky the fix is, there are many -ati users who has got perfectly stable systems and we cannot risk introducing regressions on their systems (there is a very strict testing process for shipping updates, you can read about it under topic "SRU" in the wiki).

Long term there are good news for bugs like this one though because the -ati / -radeon writer that is currently being developed has a special debugging facility that allows the GPU command buffer to be dumped to a file. If we just had that feature today, this bug would be much much simpler to understand because then we could see exactly the set of commands that were sent to the GPU (that made it stop responding). The problem today is that today the only thing we can see is that the driver is waiting for the GPU to finish "some task" that never finishes but we can't see what that task is. When compiz is ON this manifests itself as an ioctl() call that never returns and when compiz is OFF it manifests itself as very high CPU activity in xorg itself but that only happens because what xorg is doing is that it's repeatedly asked the GPU "dude, are you done yet?". It just keeps asking again again but the GPU is never idle so it can never start to process that next request. Even though the symptoms look very different, it's actually the same bug. Because of the lack of a good debugging facility, this type of bug is actually quite hard to fix/narrow down and I won't hold my breath for a fix (even though I hope for one just as much as you do).

Maybe in the meantime you can run with "DRI" "off" or "NoAccel" "true" or something like that?

@juhuu, after talking to the devs and after more careful investigation I've changed my mind. I think you're right juhuu. I don't think this is a dup of 371279 after all.

You can add me to the list too. I'm having the same issues. I tried adding:

 Option "AccelMethod" "XAA"

but it didn't stop the freezes and in fact also added the black screen issue.

Why is this bug tagged “compiz”? The original bug occured when using KWin, not Compiz, and several people reported that they have no problems with Compiz.

I agree with Wolfgang. I don't even have compiz installed. Tags should probably be xorg, ati and maybe kwin (although I doubt that kwin is the perpetrator here..).

Kernel upgrade to 2.6.29-4 seems to have worked for me. So far so good

Wolfgang Jeltsch, this bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? If so, could you please test for this with the latest development release of Ubuntu? ISO images are available from http://cdimage.ubuntu.com/daily-live/current/ .

If it remains an issue, could you please run the following command in the development release from a Terminal (Applications->Accessories->Terminal), as it will automatically gather and attach updated debug information to this report:

apport-collect -p xserver-xorg-video-ati REPLACE-WITH-BUG-NUMBER

Please note, given that the information from the prior release is already available, doing this on a release prior to the development one would not be helpful.

Thank you for your understanding.

Helpful bug reporting tips:
https://wiki.ubuntu.com/ReportingBugs

[Expired for xserver-xorg-video-ati (Ubuntu) because there has been no activity for 60 days.]