The modetest program with -v argument should flip back and forth between a color test pattern and a solid grey screen. Running just "./modetest" said I was running 1600x1200 @75hz. I ran "./modetest -s 7:1600x1200 -v" and "./modetest -s 7:640x480 -v", both gave me a lot of garbage. I ran "./modetest -s 7:1600x1200@75 -v", and got an almost steady display alternating vertically between bars of the color pattern and the solid grey. Very roughly 20 rows of pixels each. Later I ran just "./modetest -s 7:1600x1200@75" and got a kernel panic. I got a photo, but I need to find my CF reader. It included: ...RIP... nouveau_finish_page_flip+0xae/0x1a0 [nouveau] I was using an NV50 family card, G92 [GeForce 8800 GT]. I believe this is related to the corruption I'm occasionally seeing in wayland's DRM output. Built kernel from nouveau kernel repo, last commit: commit 82eeca9a994c9676e3ab478285d02b9474d70cba CommitDate: Thu Feb 17 11:11:28 2011 +1000 libdrm freshly rebuilt from git: commit a697fb6acad7992c3d23bb6a663663694782eb7b CommitDate: Thu Feb 17 20:40:30 2011 +0100
Created attachment 43543 [details] Kernel panic Original is higher resolution if there's anything you can't read in this.
The very next commit in the tree (8e645575d469bf08c9d5d98a101ef4cfce6a9180 "drm/nouveau: Fix pageflip event") has a fix related to page flipping, could you please try it.
Sorry I failed to mention, I had applied that "drm/nouveau: Fix pageflip event" patch manually, before it was committed. Without it the screen never update in wayland (DRM), staying frozen on the first frame. I expect modetest would have done the same.
This was apparently fixed at some point before Ubuntu kernel 3.0.0-17-generic was packaged. Any idea which commit it was?
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.