Bug#550977:
Hello, I can confirm that file-system corruption happens with 2.6.31-1-686 too. It may be subjective but it seems to me that it happens with 2.6.31 less often then it did with 2.6.30. If no file-system corruption occurs, then the apps which I tested (googleearth, Max Payne) sometimes crashing with: drmRadeonCmdBuffer: -22. Kernel failed to parse or rejected command stream very similar to bug #551393 or they crash and all other applications which I try to start after that make a Segmentation fault. Greetings, Andy -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#550977:
As another data point, does this also happen after a fresh boot, or only after a suspend/resume cycle? The problem also occurs after a reboot. I tried it twice this evening and both times the filesystem got corrupted. Greetings, Andy -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#268611: unsubscribe
Joshua Kwan wrote: Christoph Hellwig wrote: Just download the .diff.gz from kernel-source instead and run the included prune-non-free script. kernel-patch-debian is a useless package. The point of kernel-patch-debian, as far as building our kernels go, is: Suppose a kernel-image source package build-depends on kernel-tree-2.4.27-4. This is provided by kernel-tree-2.4.27, which in turn depends on kernel-source-2.4.27 and kernel-patch-debian-2.4.27, but it is version 2.4.27-5 in unstable. And let's say -5 introduced some fix that prevents kernels meant to build against -4 from building against it. Herbert's kernel-image build system will call /usr/src/kernel-patches/all/$VERSION/apply/debian 2.4.27-4 based on some wacky sed magic applied to the kernel-tree-2.4.27-4 build-dependency, and the extracted kernel tree will appear just as it does in a kernel-source-2.4.27 tarball from 2.4.27-4. (I hope that made sense..) Yes, it's useless in the eyes of an end user but prevents spurious build failures due to inconsistencies between kernel-source package versions. *shrug*