CCing jbarnes who is the guru on this. Hopefully he can shed some light. Bryce
On Thu, Feb 18, 2010 at 10:21:14PM +0100, Geir Ove Myhr wrote: > Hi xorgers, > > We used to ask for batchbuffer dumps for GPU hang bugs (freezes) on > intel in order to help upstream figure out the cause. It turns out > that these are currently not useful for upstream anymore with Lucid, > since kernel 2.6.32 automatically resets the GPU after a hang is > detected. Chris Wilson has an experimental kernel patch [1] that will > save the state of the batchbuffer to > /sys/kernel/debug/dri/.../i915_error_state [2]. I have tried to > compile 2.6.33-rc8 with the patch, but after following the > instructions on [3] the kernel doesn't boot with my Lucid. Would it be > possible for someone with more kernel knowledge than me to build a > recent kernel with the patch and put it in a semi-official place so > that we can ask people with freeze bugs to record the batchbuffer with > that kernel? > > I have updated https://wiki.ubuntu.com/X/Troubleshooting/Freeze to > reflect the situation. > > Btw, does anybody know more about how the GPU resetting works? What > happens after a GPU hang is detected and the GPU is reset? Does xorg > need to be restarted or should it in theory recover automatically? > > [1]: http://lists.freedesktop.org/archives/intel-gfx/2010-February/005894.html > [2]: http://lists.freedesktop.org/archives/intel-gfx/2010-February/005810.html > [3]: https://wiki.ubuntu.com/KernelTeam/GitKernelBuild > > Geir Ove > > -- > Ubuntu-x mailing list > [email protected] > Modify settings or unsubscribe at: > https://lists.ubuntu.com/mailman/listinfo/ubuntu-x -- Ubuntu-x mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-x
