| Summary: | Screen corruption when using Xorg 6.7.99.903, xcompmgr -cf, nvidia twinview on amd64 architecture | ||
|---|---|---|---|
| Product: | xorg | Reporter: | Marcus D. Hanwell <cryos> |
| Component: | Driver/nVidia (proprietary) | Assignee: | Aaron Plattner <aplattner> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | major | ||
| Priority: | high | CC: | bjdouma, framerate, melancholia.abyss, svmaris |
| Version: | unspecified | ||
| Hardware: | x86 (IA32) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Marcus D. Hanwell
2004-08-31 18:16:55 UTC
I have this same screen corruption problem when running xcompmgr, I noticed that if I turn off twinview, it doesn't happen - seems related somehow Just to add that the same problems are observed in xorg-x11-6.7.99.904 too - in fact I can see no change at all with my particular set up. I have the same problem with screen corruption, but not with redrawing windows. I have XOrg 6.8.0, xcompmgr from CVS and nVidia driver version 6111, running on ia32. I'm also using TwinView. Here's a screenshot taken soon after starting xcompmgr: http://ion.yi.org/~ion/tmp/xcompmgr-bug (PNG, 2048x768, 285k) Judging from both screenshots, it seems like the corrupted data fills the whole screen vertically, and is always 1792 pixels wide. This bug also exists for non standard aspect ratios on single display pcs. i have the same corruption on my dell i8600 (Pentium-M) with 1920x1200 (16:10) display. not only amd64.. happens on all twinview's I'd assume, or possibly all resolutions that aren't standard.. (In reply to comment #5) > not only amd64.. happens on all twinview's I'd assume, or possibly all > resolutions that aren't standard.. p.s. Im running xorg 6.8.0-r1 and fluxbox .9.10-r3 on kernel 2.6.8-r3 gentoo. This seems to be due to large windows (including the root window). When running my resolution at 2048x1536 i get this background corruption on my root window. And if i maximise a window, that too gets corrupted. I have since reduced my resolution to 1600x1200 and no longer get any issues. My guess is it is due to the size of the image being drawn offscreen. (please excuse me from not being too precise in my terminology). I no longer get any corruption, though I would like to be able to run my screen at 2048x1536 again as its nice to have the desktop space ;) (In reply to comment #7) > This seems to be due to large windows (including the root window). That's what my hunch was. I have the same problem (actually a bit compounded) with a virtual 2000x1600. That's with vtwm window manager (yaaay!), using nvidia 6629 with most of the good options turned on (RendelAccel to be specific). The nv_drv doesn't display these problems, I might add. bjd (In reply to comment #7) > This seems to be due to large windows (including the root window). That's what my hunch was. I have the same problem (actually a bit compounded) with a virtual 2000x1600. That's with vtwm window manager (yaaay!), using nvidia 6629 with most of the good options turned on (RendelAccel to be specific). The nv_drv doesn't display these problems, I might add. bjd (In reply to comment #9) > (In reply to comment #7) > > This seems to be due to large windows (including the root window). > > That's what my hunch was. > > I have the same problem (actually a bit compounded) with a virtual > 2000x1600. That's with vtwm window manager (yaaay!), using nvidia > 6629 with most of the good options turned on (RendelAccel to be > specific). > > The nv_drv doesn't display these problems, I might add. > > bjd Same behaviour just confirmed with fvwm. Again, this is with a virtual screen of 2000x1600. bjd I've found a bug in the xorg X server. Some programms in Linux read the wallpaper bto generate transparency, for example idesk or gkrellm or pypanel. Since I updated to xorg 6.8, they can't read the set up wallpaper and draw the original background of xorg. <a href="http://www.makoe.com/images/xorg_bug.jpg">Here is an example</a> I can confirm this behaviour on a Dell Inspiron 9300 - NVIDIA GeForce 6800 Go with a res. of 1920x1200, running Ubuntu Linux (Hoary) with X.org 6.8.2 and NVidia binary drivers 1.0.7167. Corruption occurs on Gnome-startup (when xcompmgr starts), because of the big root-window. I can 'wipe' the artifacts out with other (smaller) windows. Everytime a window gets a size bigger than, say, 1700x1000 the corruption occurs again and disappears when I make the window smaller. It looks the same as others have posted in their screenshots. Since there are no updates on this bug after almost 3 months, what is the current status? (In reply to comment #11) > I've found a bug in the xorg X server. > Some programms in Linux read the wallpaper bto generate transparency, for > example idesk or gkrellm or pypanel. Since I updated to xorg 6.8, they can't > read the set up wallpaper and draw the original background of xorg. > > <a href="http://www.makoe.com/images/xorg_bug.jpg">Here is an example</a> > > AFAIK, this is not related to the bug presented here. I am also seeing display corruption, matching the description in the other comments; I'm using kompmgr, the NVIDIA display drivers 1.0-7174 (with TwinView active), and Gentoo's CVS snapshot labeled "xorg-6.8.99.5", on amd64. Im using xorg-x11-6.8.99.15-r3 on Gentoo. Fluxbox is my window manager and I am still getting the corruption in twinview mode. It actually has been this way since the composite extension was introduced. Out of curiosity this morning I was messing around and I disabled RenderAccel in my xorg.conf and there is no corruption. This left me with a very slow desktop as Im sure you all would know. At any rate, this is somehow related to RenderAccel which would make this an NVidia bug. My installed version of the NVidia driver is 1.0-7174. This issue has been tracked down and will be fixed in a future driver release. That is great to hear. I would be more than happy to test out the fixed driver once it is available. This should be fixed with the 1.0-8178 driver, available at http://www.nvidia.com/object/unix. Marcus, please test and close or reopen this bug as appropriate. (In reply to comment #18) > This should be fixed with the 1.0-8178 driver, available at > http://www.nvidia.com/object/unix. Marcus, please test and close or reopen this > bug as appropriate. I did a quick install and de-install (this is under Xorg 6.8.2) and am finding that there is no difference in the way it messes things up from what the earlier releases did. bj (In reply to comment #19) > I did a quick install and de-install (this is under Xorg 6.8.2) and am > finding that there is no difference in the way it messes things up from > what the earlier releases did. Bauke, can you please attach a screenshot of the corruption with 1.0-8178 and an nvidia-bug-report.log? Thanks! Aaron this appears to be fixed for me here. I am running TwinView as before, new GPU (6600GT, PCI-E now) with a 2.6.14 kernel and xorg-x11-6.8.99.15. I am going to try upgrading to the new modular xorg release on this system soon, but I expect it will continue to function. It is a Gentoo Linux system. Please let me know if you need any more information. I am using KDE 3.5 and the built in translucency support now with RenderAccel on. I would be happy to provide any more information you might need. Thanks for fixing it! (In reply to comment #21) > Aaron this appears to be fixed for me here. Great! If it works for you, then I think it's safe to close this bug now. Thanks for testing, everybody! I'm sorry for not replying sooner, but as far as I'm concerned this bug isn't fixed at all with NVidia's 1.0-8178 driver. Am I understanding correctly that this does NOT also hinge on use of latest Xorg 6.9/7.0? Anyway, I've placed three root screenshots at http://www.xs4all.nl/~bjdouma img-root-20051224-095423.png: right after starting /usr/x11/bin/xcompmgr -f -D 8 -I 0.025 -O 0.025 -l -5 -t -5 -c -r 8 )img-root-20051224-095504.png: right after iconifying and de-iconifying the XTERM-0 xterm. img-root-20051224-095526.png: right after iconifying the emacs window. nvidia-bug-report.log: the log that goes along with these I am using a virtual 2000x1600. Window manager is fvwm. I'll be glad to provide any other details. bj Bauke, Based on your screenshots, your problem looks different from the one reported by Marcus. Please file a new bug and assign it to me. Good things to include in the new bug are the screenshots and nvidia-bug-report.log you posted and answers to the following questions: Does your problem occur if you use a screen narrower than 1700 pixels or a different window manager? Also, if you have KDE installed, does it work with kompmgr, rather than xcompmgr? Thanks! The problem fixed in 1.0-8178 does not depend on the version of X.org being used. Moving back to RESOLVED FIXED. *** Bug 5320 has been marked as a duplicate of this bug. *** |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
