I have managed to work my way back up the failure chain, to the point where GDB can stop the server before the mouse moves. The failure is associated with a motion event of {0,0} and is_absolute of 1.
(gdb) bt #0 xf86PostMotionEventP (device=0x84a6308, is_absolute=1, first_valuator=0, num_valuators=2, valuators=0xbfc6af7c) at ../../../../hw/xfree86/common/xf86Xinput.c:741 #1 0xb7a6a263 in ?? () from /usr/lib/xorg/modules/input//evdev_drv.so #2 0x084a6308 in ?? () #3 0x00000001 in ?? () #4 0x00000000 in ?? () (gdb) p *valuat...@num_valuators $77 = {0, 0} That would definitely send the pointer to {0,0}. Unfortunately, I'm using a mouse. And, of course, the absolute motion event didn't show up in my original traces because I never bothered to put a trace on that call, because it was absolute motion. Back to debugging evdev. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org