Bug 549226

Created attachment 323742 [details]
Next the cursor

http://img148.imageshack.us/img148/3341/artef1.jpg

Next the cursor

http://img88.imageshack.us/img88/7162/artef2.jpg

Bands after closing the window

http://img91.imageshack.us/img91/9875/artef3.jpg

Next the cursor

http://img88.imageshack.us/img88/2378/artef4.jpg

Bands after moving the window

http://img88.imageshack.us/img88/5851/artef5.jpg

Next the cursor

Actual Results:  
After closing the windows and menus remain the color bands of the window. Also 
there are traces of the cursor employment

The same is happening for me on my Radeon X1950 (radeonhd doesnt work but thats another bug) running 11.2 RC1 . Except for these artifacts everything works with this driver

The same problem with X1300 (recognized as X1600 which is another story) with radeonhd on 11.2 RC1.

Reassigning during vacation.

Could you please supply a log file?

Could you explain what attracts you to a log file and where to find it in the system?

# dmesg                                                   
[    0.000000] Linux version 2.6.31.3-1-desktop (geeko@buildhost) (gcc version 4.4.1 [gcc-4_4-branch revision 150839] (SUSE Linux) ) #1 SMP PREEMPT 2009-10-08 00:27:25 +0200                                                                   

[   25.215543] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[   30.219229] BIOS EDD facility v0.16 2004-Jun-25, 3 devices found
[   32.323964] ACPI: PCI Interrupt Link [APC5] enabled at IRQ 16
[   32.323976]   alloc irq_desc for 16 on node 0
[   32.323979]   alloc kstat_irqs on node 0
[   32.323987] pci 0000:01:00.0: PCI INT A -> Link[APC5] -> GSI 16 (level, low) -> IRQ 16
[   32.399586] [drm] Initialized drm 1.1.0 20060810
[   32.425269] [drm] Initialized radeon 1.31.0 20080528 for 0000:01:00.0 on minor 0
[   32.693299] agpgart-amd64 0000:00:00.0: AGP 3.0 bridge
[   32.693322] agpgart-amd64 0000:00:00.0: putting AGP V3 device into 8x mode
[   32.693357] pci 0000:01:00.0: putting AGP V3 device into 8x mode
[   32.938580] [drm] Setting GART location based on new memory map
[   32.938592] [drm] Loading R300 Microcode
[   32.938638] [drm] Num pipes: 2
[   32.938648] [drm] writeback test succeeded in 1 usecs
[   35.953022] eth0: no IPv6 routers present

(In reply to comment #8)
> Could you explain what attracts you to a log file and where to find it in the
> system?

For X.Org this is /var/log/Xorg.0.log. :-)

Created attachment 324529 [details]
Xorg.0.log

Need more information?

This bug is present even in RC2

OpenGL vendor string: DRI R300 Project                                        
OpenGL renderer string: Mesa DRI R300 (R350 4E48) 20090101 AGP 8x x86/MMX+/3DNow!+/SSE2 TCL                                                                                       
OpenGL version string: 1.5 Mesa 7.6                   

>>Mesa 7.6 is released. This is a new feature release. Those especially
>>concerned about stability may want to wait for the follow-on 7.6.1 bug-fix release.

http://www.mesa3d.org/

Is not this the case?

These artifacts seem to be present on all radeon generations regardless of the driver. It looks like an Mesa + AIGLX issue with desktop effects.

This bug still exists in the final 11.2 release.

I also have this problem. Using 11.2 64bit release with an ATI X700 mobile.

I briefly tested a Fedora12(kde) and kubuntu livecd, and neither presents these problems. 

All three use Mesa 7.6. But while opensuse uses ati driver 6.12.4, fedora uses ati driver 6.13 (20091006git) and kubuntu uses 6.12.99 (29092009git), so it seems to be corrected in future driver versions. 

Is there a "simple" way we can get a more up to date ati driver? ... desktop effects are completely unusable as it is.

regards,
Carlos

(In reply to comment #16)
> All three use Mesa 7.6. But while opensuse uses ati driver 6.12.4, fedora uses
> ati driver 6.13 (20091006git) and kubuntu uses 6.12.99 (29092009git), so it
> seems to be corrected in future driver versions. 
> 
> Is there a "simple" way we can get a more up to date ati driver? ... desktop
> effects are completely unusable as it is.

I'm afraid it isn't. Most likely Fedora 12 and Ubuntu already use the bleeding
edge radeon stuff including KMS. See http://airlied.livejournal.com/66958.html
for more details.

*** Bug 556642 has been marked as a duplicate of this bug. ***

Also same problem with Ati X600 Mobility on HP nc8230 and openSUSE11.2.

I have the same problem with Ati x1400 Mobility on a Dell Inspiron 6400 and openSUSE11.2.

Since Mesa 7.7 from the Xorg-repository this problem has gone for me. As I have installed the Kernel 2.6.32RC8 from Kernel-Factory-repo, I don't know exactly if it is enough to install the packages from xorg-repo in order to fix this.

So could perhaps someone test this with the 11.2 standard kernel?

I confirm that updating Mesa library solves the problem of the stripes even with standard 11.2 kernel (HP Compaq 6715b, Radeon 1250, R300). The only problem still remaining is the occasional flickering of the screen with kwin desktop effects switched on.

I've also tested the files from the Xorg repo. Although the stripes do disappear i have some "brief" corruption while loging in and after a few seconds the whole pc frezes - have to poweroff/poweron (radeon X700 mobile).

regards,
Carlos

It is out hope for a patch with an updated open source ATI driver and updated MESA3d??

For me it's gone on all my PCs/Notebooks with radeon using Mesa 7.7.

So I think Carlos' freeze (comment 23) might be another bug.

So when will be available maintenance update to Mesa 7.6.1 or higher?

(In reply to comment #26)
> So when will be available maintenance update to Mesa 7.6.1 or higher?

If I understood correctly the issue is not fixed before Mesa 7.7. Updating to a
development release (which 7.7 is) - or any other version upgrade - would be against our maintenance update policy.

In fact, the issue is not fully solved even in Mesa 7.7, only the occurrence of the strip flickers is significantly rarer. (See comment 22).

"would be against our maintenance update policy"

True.

However, not seeing the bloody cursor when typing text in Firefox or Thunderbird is a big stinky business.

Resolving this would be "providing bug fixes for major bugs"--also a nice policy we all appreciate. :-)

Do we have the problem of the unstoppable force against the immovable object?

Cannot the fix be backported from Mesa 7.7 to Mesa 7.6.1?

So I build and install opensource driver xf86-video-ati-6.12.4 and mesa3d 7.6.1 and I have passed this bug. 3D and compósiting in KDE 4.3.5 works

http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/snapshot/xf86-video-ati-6.12.4.tar.bz2

http://cgit.freedesktop.org/mesa/mesa/snapshot/mesa_7_6_1.tar.bz2

Discovers that he became ill working suspend to disk set
mesa 7.8 devel (((

Is this still an issue with openSUSE 11.3?

There is a problem opensuse 11.2

> > Is this still an issue with openSUSE 11.3?
> There is a problem opensuse 11.2

???

I can confirm that it doesn't happen in 11.3, but I understand that it still happens in 11.2.

Thanks. So let's close it as fixed for openSUSE 11.3.