Created attachment 28339 [details] Xorg log [forwarded from Debian bug#538442, submitted by Bastian Blank <waldi@debian.org>] "The new intel driver likes to report errors in the hardware state after some time running. | (WW) intel(0): ESR is 0x00000011, page table error, instruction error | (WW) intel(0): PGTBL_ER is 0x00000010, display A pte | (WW) intel(0): Existing errors found in hardware state. Using Xv just makes it happen faster and completely disables the output (no log currently). Without Xv the effects includes - glyphs are replaced by rectangles, - background image looks like memory corruption and - wrong colors." kernel 2.6.30, libdrm 2.4.12, xserver 1.6.2, xf86-video-intel 2.8.0 The hardware is asus eeepc. X log attached. This is using a PAE kernel, which means no GEM. I understand this is a compat path, but it seems there's some issue when using the fake bufmgr and the effects seem quite bad. I asked Bastian to try NoAccel, will report back.
(In reply to comment #0) > I asked Bastian to try NoAccel, will report back. ignore that part, I had forgotten that NoAccel was removed from 2.8.0. sigh.
I'm not sure if we need fix this path, since GEM has been enabled on PAE kernel now (bug#17993)
> --- Comment #2 from Gordon Jin <gordon.jin@intel.com> 2009-08-04 18:56:59 PST --- > I'm not sure if we need fix this path, since GEM has been enabled on PAE kernel > now (bug#17993) aiui new drivers are still supposed to have some support for non-gem kernels, if only to allow an upgrade path...
The PAE issue is long fixed, so there should be no reason to use broken compat paths any more.
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.