[rv350 9550] Graphical artifacts present when resuming from sleep.

[Derek White]

No idea what is causing the problem, but when I resume from sleep (gnome-power-preferences calls it sleep, but the power icon calls it suspend...wtf???) ['S1&S3' selected in BIOS, computer completely off, keyboard/mouse will not wake, have to use power button to resume - not hibernation, hibernation is disabled, computer comes right up to locked screen, no booting...no idea what to call this type] the upper & lower gnome menus have artifacts around them, as well as any pop-up menus (i.e. Applications, Places, right click on desktop/firefox/anywhere). See attached picture.

Ubuntu 8.10 i386-generic
Default graphics drivers for ATI Radeon 9550 (RV350)
   xserver-xorg-video-radeon 1:6.9.0+git20082003.f9826a56-0ubuntu2
   $ lsmod | grep radeon
      radeon 147616 2
      drm 86056 3 radeon
Default appearance settings (Visual Effects: Normal)
   compiz/compiz-core/compiz-gnome/compiz-plugins/compiz-wrapper -- 1:0.7.8-0ubuntu4
   compizconfig-backend-gconf -- 0.7.8-0ubuntu1
   compiz-fusion-plugins-extra/compiz-fusion-plugins-main/libcompizconfig0 -- 0.7.8-0ubuntu2
   $ pgrep comp
      16462 compiz
      16528 compiz.real
      16538 compiz-decorato
$ cat xorg.conf
$ [AKA: file exists, but is empty]

Artifacts can be removed immediately by logging out & back in.
Also can be removed by using the gnome menus (i.e. Applications/Places/System) over and over. (While writing this they disappeared.)

Not high priority, but very ugly and messy. Kinda seems like the shadowing is all artifact-ed out.

Notes on the picture: The menu is usually visible on the top, the weird static is a common ati gfx problem that occurred when I snapped the picture. The problem I am reporting is the messed up border seen around the bottom bar and around the top bar. I can get better picture if needed, including the right-click menu and other menus.

[TomasHnyk (sup)]

I have the same chipset (ATI Radeon 9700, but it is rv350) and I am seeing the same artifacts as you. Thus marking confirmed.
There is a similar bug (https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/272969) , though I do not know if it is a duplicate. But there is definitely a workaround that fixes this for me: Option "AccelMethod" "exa"
to the device section of /etc/X11/xorg.conf.

Please try if it fixes the issue for you. BTW:if it does not, the quickest way to get rid of the artifacts is to get to another virtual console and then back (alt+ctr+f6, then alt+ctrl+f7).

Also, I do not know if it matters that I do have compiz running - I can of course test.

I hope I got the assignation and affected package right. If not, please reassign.

[TomasHnyk (sup)]

After running for several hours with:
Option "AccelMethod" "exa"
in my xorg.conf I think it hinders the performance and when I disabled compiz, it was impossible to enable it back. Without it, I could turn it on alright.

[Tormod Volden (tormodvolden)]

> Artifacts can be removed immediately by logging out & back in.

Is it enough to switch console? Ctrl-Alt-F1 and then Alt-F7 back.

[TomasHnyk (sup)]

For me it is enough.

[Derek White (d-man97)]

Switching console seems to work.

Also, with the jockey version of fglrx, the problem is not there.

[Bryce Harrington (bryce)]

This was a bug in the XAA acceleration architecture, which is resolved by specifying AccelMethod EXA in your xorg.conf. Note that in Intrepid EXA has some performance regression compared with XAA, so it may or may not be worth switching. For Jaunty, EXA performance is much improved, and we moved to EXA by default.

If by chance you still reproduce the problem in Jaunty, feel free to reopen the bug; please attach a new Xorg.0.log in this case.

[Dominik (dominalien)]

I am having a similar problem when resuming from hibernation in Jaunty.

I posted my comments in a similar bug: https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/272969

Key points:

It's Jaunty using radeon driver which comes with it.
This is on r200 (Mobility Radeon 9200).
Switching consoles worsens the problem.
Logging out and back in solves the problem.
Resuming from suspend on this machine (Acer Ferrari 3000) results in a hang (bug: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/339336 )

Should I reopen this bug? Or should I open a new one?

Attaching my Xorg.0.log file (grabbed right after resuming from hibernation). What other information should I provide?