Created attachment 278247 [details] dmesg output I just had a crash using GoogleEarth. First parts of the ground were covered not with the expected ground images, but instead some contained what looked like random noise, while other contained transformed versions of images that appeared on screen before, i.e. content from bubble boxes for e.g. fotos. After changing the view of this thing for a while, the screen turned black and printed a kernel error message: "unable to handle kernel NULL pointer dereference". Surprisingly, the mouse cursor was still visible in front of this error message background, and I could even ssh into the box and save the dmesg output, which I've attached. Notice the kernel error message in lines 2162 through 2210, timestamps 29702.140686 through 29702.550133. Packages which might be involved in this: sys-kernel/gentoo-sources-2.6.39-r1 x11-drivers/xf86-video-nouveau-0.0.16_pre20110323 sci-geosciences/googleearth-6.0.3.2197 x11-base/xorg-server-1.10.2 x11-libs/libdrm-2.4.26 I'm beginning to wonder whether nouveau is any more stable than the proprietary drivers. Both kinds cause a LOT of trouble for me lately. This is becoming increasingly annoying. In case there is any relation: the last nouveau bug I encountered was bug #371841.
This appears to be a general problem with TTM but there seems to be no confirmation on what causes it. In one bug report it is suggested to increase AGP aperture in the BIOS (if your card is an AGP card).
I have problem with crashing GoogleEarth at statrup too. in crashlog have only this: Major Version 6 Minor Version 0 Build Number 0003 Build Date May 17 2011 Build Time 00:40:40 OS Type 3 OS Major Version 2 OS Minor Version 6 OS Build Version 37 OS Patch Version 0 Crash Signal 11 Crash Time 1312637811 Up Time 0.881362 Stacktrace from glibc: ./libgoogleearth_free.so(+0xab953)[0xb76bb953] ./libgoogleearth_free.so(+0xabad3)[0xb76bbad3] [0xb7729400] I try with enableTyps=false but without any result. With gdb evrething endded with: . . . [Thread 0xabbf2b70 (LWP 5329) exited] [Thread 0xacbf4b70 (LWP 5327) exited] [Thread 0xad3f5b70 (LWP 5326) exited] [Thread 0xaef92b70 (LWP 5323) exited] [Thread 0xae791b70 (LWP 5324) exited] [New Thread 0xad3f5b70 (LWP 5335)] [New Thread 0xacbf4b70 (LWP 5336)] [New Thread 0xabbf2b70 (LWP 5337)] [New Thread 0xa7b39b70 (LWP 5338)] [New Thread 0xa7338b70 (LWP 5339)] [New Thread 0xa6b37b70 (LWP 5340)] [New Thread 0xa6336b70 (LWP 5341)] [New Thread 0xa5b35b70 (LWP 5342)] [Thread 0xacbf4b70 (LWP 5336) exited] [New Thread 0xa4cefb70 (LWP 5343)] [New Thread 0xa44eeb70 (LWP 5344)] [New Thread 0xa3cedb70 (LWP 5345)] [New Thread 0xa34ecb70 (LWP 5346)] [New Thread 0xa2cebb70 (LWP 5347)] [New Thread 0xa24eab70 (LWP 5348)] [New Thread 0xa1ce9b70 (LWP 5349)] [Thread 0xb05edb70 (LWP 5318) exited] [Thread 0xb0deeb70 (LWP 5317) exited] [Thread 0xb15efb70 (LWP 5316) exited] Program received signal SIGSEGV, Segmentation fault. 0x00000000 in ?? () The system is 32bit. Videocard is NVidia 460 PCI with 1GB DDR Have anybody something about this problem.
Could reproduce twice with: media-libs/mesa-7.11 sci-geosciences/googleearth-6.0.3.2197 sys-kernel/gentoo-sources-3.0.3 x11-drivers/xf86-video-nouveau-0.0.16_pre20110711 x11-libs/libdrm-2.4.26 I found out that surprisingly the X session can be recovered: switching from the text mode screen with the kernel error message in white on black to any VT and then back to VT11 will restore X to working order, except for googleearth which has to be killed. Haven't encountered any broken rendering prior to the crash this time around, so that might have been coincidence in comment 0.
Does this crash still occur with later kernels?
(In reply to comment #4) > Does this crash still occur with later kernels? Haven't encountered it in a while. But as this always was a sproradic thing, we can never be sure. Feel free to close, as long as I may reopen once this turns up again.
Please feel free to reopen if this happens again. I'll put invalid for now for lack of a better choice.