Distribution: Gentoo 2005.0 Hardware Environment: i686 AMD Athlon(tm) 64 Processor 3400+ Software Environment: gcc-3.4.4, glibc-2.3.5-r0, Xorg-6.8.2 Problem Description: When running under kernel 2.6.12.2, the Xorg wacom driver locates the /dev/input/eventX devices and correctly interprets the X options to make the mouse/puck run in relative mode, such that picking up the device and replacing it on the other side of the tablet does not cause the cursor to move at all. Upon choosing the 2.6.13_rc1 kernel under grub, X boots fine and provides no errors, it reports that the mouse has been told to run in relative mode, but the cursor now moves to the absolute position on the screen when the puck is removed and replaced on the other side of the tablet. The 2.6.13-rc1 config was created by copying the 2.6.12.2 config and running make oldconfig and altering only the new options. The linuxwacom driver (both kernel and X version) were present during this problem. Since the linuxwacom Xorg driver currently doesn't compile for dlloader, it wasn't possible to test whether this worked or not. This was reported http://bugs.gentoo.org/show_bug.cgi?id=99178, where it was suggested that I file a kernel bug. Steps to reproduce: Boot kernel-2.6.12.2 or lower. Setup the mouse to relative mode and check that it works as expected. Reboot the machine in kernel-2.6.13-rc1 (or higher up to _rc3 so far) Remove and replace the mouse on the tablet. If there's any further information I can provide that might help, or tests I can run etc., please post them and I'll be happy to give them a try. Thanks...
Is this problem still present in 2.6.13-rc4? Were you able to find a workaround?
The problem still occurs with 2.6.13_rc4 (and I've updated the summary accordingly). I've now been able to try the linuxwacom driver over the one included with Xorg since the gentoo developers very kindly included dlloader support in a recent ebuild, and the problem still occurs both with and without it present. As previously, returning to 2.6.12.2 cures the problem. I haven't found any other workaround yet. Assuming it's related to the event driver, does anyone know exactly what changed between versions in that area? As always if anyone needs me to run any tests or try anything for them, I'm more than willing, just let me know what I should do. Thanks...
hm. Dmitry, Vojtech: is this related to the in-kernel Wacom driver or is it a general input driver core problem?
I think it was caused by the latest Wcom driver update from Ping Cheng that I picked up from -mm. I was hoping we would comment - he was CCed on th original e-mail.
I'm now in contact with Ping and we're trying to figure out the problem. So far we've determined that replacing the wacom drivers from the 2.6.13 sources with ones from the 2.6.11 sources still shows the same symptoms. I'll also report back anything else we find here...
From Mike's last comments (#5), I don't think it is a problem of my latest wacom.c. And, I don't think it is the new kernel's issue either. I feel it is probably a driver setup or X configuration file issue. Mike is doing more testing there. We'll keep you informed.
Ok, I'll mark this as INVALID for now. If you all determine that this is a kernel issue, please reopen the bug.
Created attachment 5630 [details] fixes this bug and a Cintiq 21UX bug It turns out that the code does have a bug for Graphire. Attached patch fixes the bug and a Cintiq 21UX bug as well.
Please email the patch in the normal manner if it's ready for applying.
Got it, I'll forward it on now, thanks.