| Summary: | touching a button on Intuos5 makes a "silent" cursor jump to top-left corner | ||
|---|---|---|---|
| Product: | xorg | Reporter: | Timo Aaltonen <tjaalton> |
| Component: | Input/wacom | Assignee: | Jason Gerecke <killertofu> |
| Status: | RESOLVED FIXED | QA Contact: | Xorg Project Team <xorg-team> |
| Severity: | normal | ||
| Priority: | medium | CC: | florian, peter.hutterer |
| Version: | git | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Timo Aaltonen
2012-08-30 08:00:37 UTC
by "silent" I mean the cursor doesn't seem to jump there, but gimp shows these streaks..
xev shows this even when that happens:
KeymapNotify event, serial 42, synthetic NO, window 0x0,
keys: 4294967235 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
still the same with kernel 3.6-rc5, -wacom 0.17. After some poking around, I think I've tracked down the cause of this behavior. When touching (but not pressing) a button, the kernel driver mostly ignores the hardware event it receives. The out-of-prox logic is triggered, and we send an event indicating the pad has left proximity. The problem is that since the pad never *entered* proximity, it can leave the X driver a little confused. When a new packet comes in from the kernel, the X driver uses the serial number to find the appropriate "channel" to store event data in. It then updates the most-recent information in that channel with the packet data and processes the result as an event. An excess out-of-prox event starts from an empty state, causing the driver to ask the kernel for the tool that is currently in proximity. If it finds something, an empty event is sent for that tool, causing a momentary jump to the top-left corner. There are two ways I see to fix this bug: a) have the kernel send an (empty) in-prox event for button touches so the X driver doesn't get confused, b) have the X driver ignore out-of-prox events for already out-of-prox tools. I could certainly use a patch for the X driver, faster to get it accepted as a stable update :) Don't mind if the kernel would get fixed later. Patches have been posted to the linxwacom-devel and linux-input mailing lists. X driver patch: https://sourceforge.net/mailarchive/forum.php?thread_name=1347664967-1976-1-git-send-email-killertofu%40gmail.com&forum_name=linuxwacom-devel Kernel driver patch: https://sourceforge.net/mailarchive/forum.php?thread_name=1347665405-2078-1-git-send-email-killertofu%40gmail.com&forum_name=linuxwacom-devel I've tested the X driver patch, and it's working great, thanks. I'll test the kernel one too (without the patched -wacom driver). Patches have been accepted upstream. A patch referencing this bug report has been merged in Linux v3.7-rc1: commit 5e056ef4ddeb6129f6bb170f2c26b8f370bbcda7 Author: Jason Gerecke <killertofu@gmail.com> Date: Mon Sep 24 09:21:31 2012 -0700 Input: wacom - mark Intuos5 pad as in-prox when touching buttons xf86-input-wacom commit: commit 0fe7f800295a4b5bf074bca798cf8024fcb174ec Author: Jason Gerecke <killertofu@gmail.com> Date: Thu Sep 13 16:59:01 2012 -0700 Ignore out-of-prox events from unknown devices already out-of-prox |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.