For some reason i got only black window instead of nice gears. It works with KMS. Kernel: drm-radeon-testing up to commit 10fd883ce384706f88554a0b08cc4d63345e7d8b Author: Dave Airlie <airlied@redhat.com> Date: Tue Apr 20 16:34:20 2010 +1000 agp/intel: put back check that we have a driver for the bridge. libdrm: commit 894c86e8400f13f1ee0dfe23ca4b5f98c32d9223 Author: Kristian Høgsberg <krh@bitplanet.net> Date: Fri Apr 9 16:33:38 2010 -0400 Revert "Fix pkgconfig includes for /usr/include/drm" xf86-video-ati: commit 2a6b409496f26da0436972b5feae6ea035dde08d Author: Owain Ainsworth <zerooa@googlemail.com> Date: Mon Apr 19 14:37:33 2010 +0100 Reference count shared driver mappings. mesa: commit 9e258fc2bd6f2b9950606a0a92bb92c8959d9efd Author: Vinson Lee <vlee@vmware.com> Date: Tue Apr 20 18:48:04 2010 -0700 gallivm: Remove unnecessary headers. xserver: commit 35d208125fa0ec78e8d694adc52886b977e789cf Author: Tiago Vignatti <tiago.vignatti@nokia.com> Date: Wed Apr 21 18:28:13 2010 +0300 xfree86: fix not reached code in fi1236 driver from i2c libpciacces: commit af2be74979aeab9a2fc4c933462e97ce70f816b6 Author: Samuel Thibault <samuel.thibault@ens-lyon.org> Date: Tue Jan 19 18:37:43 2010 +0100 libpciaccess x86 backend logs will follow .....
Created attachment 35227 [details] X log
Created attachment 35228 [details] xorg.conf Playing with colorTiling was unhelpful. Playing with resolution changes nothing. It simply doesn't work if i set DefaultDepth to 16 (for UMS). Disabling PAT support via nopat kernel option changes nothing.
Created attachment 35229 [details] dmesg
Glxgears in UMS was working (with same userspace) up to kernel commit 6d6b89bd2e316b78d668f761d380837b81fa71ef Merge: 13dda80 2507c05 Author: Linus Torvalds <torvalds@linux-foundation.org> Date: Tue Mar 2 07:55:08 2010 -0800 in mainline. I halted git-bisect because original good/bad pair was not about this bug, i wrongly assumed new kernel was unable to "fix" bug 22576. I can re-run git-bisect, but it will take time ... anyone want to beat me on this?
Created attachment 35230 [details] dmesg Sorry, original dmesg was splitted by syslog, it seems. Here is more complete version.
MTRR (under X): guest@slax:~$ cat /proc/mtrr reg00: base=0x000000000 ( 0MB), size= 512MB, count=1: write-back reg01: base=0x0d0000000 ( 3328MB), size= 128MB, count=1: write-combining reg05: base=0x0e0000000 ( 3584MB), size= 128MB, count=2: write-combining
Created attachment 35231 [details] lspci -vvvn
According to http://article.gmane.org/gmane.comp.video.dri.devel/46026 ("Regression: commit 566d84d1 breaks glxgears on r200") it was commit 566d84d172161cb6c0c4dd834c34abbac6bf7b38 (drm/radeon: r100/r200 ums: block ability for userspace app to trash 0 page and beyond) .....
At least glxgears back to normal with 2.6.35-rc1 mainline kernel ....
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.