Background corrupted after resuming from suspend

[brblueser]

Hi,

I am running Fedora 15 x86_64 with latest (270.41.19) drivers installed directly from Nvidia's own .run file. Overall it looks good -- this version seems to have solved a problem which was causing gnome-shell to crash after resuming from suspend (which is why I didn't wait for RPMFusion RPMs).

However, now when I resume from suspend everything seems functional but the background looks corrupted -- huge white rectangles all around and "mirror images" of the windows. At least changing the background seems to fix the problem, so apparently it isn't anything too serious.

[ngolian]

That sounds like the same problem I'm having on my netbook with Ion LE. It's running Debian unstable/experimental and I've had the same problem with nvidia driver versions 270.41.19, 275.09 and 275.09.04. It was OK with GNOME2, I think this only affects GNOME3.

[brblueser]

Well, at least it's not an isolated case, which increases the chances of this getting some attention ;-)

[stecklum]

Welcome, this seems to be along standing bug which obviously did not get attention by Nvidia so far...

http://www.nvnews.net/vbulletin/showthread.php?t=157417

[kwizart]

Originally Posted by brblueser

Hi,

I am running Fedora 15 x86_64 with latest (270.41.19) drivers installed directly from Nvidia's ... (which is why I didn't wait for RPMFusion RPMs).

.

The RPM Fusion repository has only version 270.41.06 (.19 will be available with the next fedora kernel currently in updates-testing). So we need to compare the same version to tell there is something wrong with the package provided.

Also some person reported that suspend doesn't work well if acpid isn't run, which could have been fixed by an update completely unrelated to the newer driver change.

[brblueser]

Originally Posted by stecklum

Welcome, this seems to be along standing bug which obviously did not get attention by Nvidia so far...

http://www.nvnews.net/vbulletin/showthread.php?t=157417

Mmmh... symptoms are similar, but I haven't seen anything like this in quite sometime (I've been using Nvidia driver + compiz since F13). It could be a regression, of course.

[brblueser]

Originally Posted by kwizart

The RPM Fusion repository has only version 270.41.06 (.19 will be available with the next fedora kernel currently in updates-testing). So we need to compare the same version to tell there is something wrong with the package provided.

Also some person reported that suspend doesn't work well if acpid isn't run, which could have been fixed by an update completely unrelated to the newer driver change.

Sure, when 270.41.19 comes from RPMFusion I will install it and repeat the tests ;-)

Regarding acpid, I had to enable it as well, since Xorg.0.log had some complaints from Nvidia driver saying it could not reach acpid (maybe the RPMs should declare acpid as a dependency).

[sandipt]

NVIDIA filed Bug ID: 816457 to track this issue for GNOME3 gnome-shell.

[brblueser]

Originally Posted by sandipt

NVIDIA filed Bug ID: 816457 to track this issue for GNOME3 gnome-shell.

Thks =) Some additional info:

• yesterday for the first time this bug didn't happen when I resumed after the computed had been suspended for a couple of hours, so it doesn't happen 100% of the times. I can't recall doing anything different, though :-(
• coincidence or not, this time when it resumed ok I was able to power off my computer -- usually it shuts down but doesn't power off, but so far I was thinking this was a kernel bug unrelated to the graphics driver (I even filed bug 712060 regarding this on Fedora's Bugzilla)

Please let me know if I can help with any additional info.

BTW: is NVIDIA's bug tracker open?