> This only contains the pile of WARNINGs and not the Oops itself. I guess > we still need a picture (and you need to change the system console logging > to make sure the debug stuff shows up to). At least at the bottom of your > past there's from drm or i915, indicating that the crucial debug info > didn't make it out before the machine died. Ok, I've set CONFIG_MESSAGE_LOGLEVEL_DEFAULT=7 to squeeze some more debug info out of the kernel. For the OOPS I still have to take a picture since kdump seems to be broken for me. http://pastebin.com/9mxRgNa2 > I just want the decoded address (which I read from the oops, if you > recompile the kernel it might have changed so please double check). I.e. > the above two lines is all the howto you need. But please use "list" > instead of "break" as Chris suggested. Gotcha! gdb) list *drm_atomic_helper_check_modeset+0x2c5 0x955 is in drm_atomic_helper_check_modeset (drivers/gpu/drm/drm_atomic_helper.c:235). 230 } 231 232 connector_state->best_encoder = new_encoder; 233 idx = drm_crtc_index(connector_state->crtc); 234 235 crtc_state = state->crtc_states[idx]; 236 crtc_state->mode_changed = true; 237 238 DRM_DEBUG_ATOMIC("[CONNECTOR:%d:%s] using [ENCODER:%d:%s] on [CRTC:%d]\n", 239 connector->base.id, (gdb) Address drm_atomic_helper_check_modeset+0x2c5/0x990 is the first line in the call trace. One more thing: When I fire up X BEFORE I connect to the docking there are no OOPSES and I can switch between DP/eDP. As soon as I kill X however my machine dies. HTH. Regards Nic _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
