Yes, I'm certain it's Elan. Any other ideas?
Yes, I'm certain it's Elan. Any other ideas?
Look in dmesg instead maybe
Code:grep -i elantech /var/log/dmesg
Nothing found there either...
Gonna have to hack at the driver then, it must be a new protocol, or at least a new magic knock if the driver isn't giving you ANY output. There haven't been any elantech patches on linux-input for a while, so if there's someone working on it now, they're probably on their own. If you're interested in working on this send me a PM and I can try to help.
Ubuntu 10.04 kernel: 2.6.32-25-generic
I have an ASUS K50ij with the elantech touchpad and it is way too sensitive. I have a hot-key to disable/enable it while I am typing, but that is not that handy when a lousy typist like me needs to mouse to make corrections.
So, I am wondering if anyone has a clean, fairly easy-to-follow fix for this yet? I would just like it to be less sensitive. The fix may be buried in this thread but my aimless wondering cannot locate the best solution. Anyone got a link to a recommended solution?
@oregonbob - Try the directions I've posted here: http://ubuntuforums.org/showthread.php?p=9175201
I'm on a Samsung QX-410 and having the same problem with the touchpad, which is recognized as "PS/2 Logitech Wheel Mouse". The pad "works" but is annoying, particularly because it's very easy to rub against the pad while typing, thus sending the mouse pointer flying all over the screen. Multitouch would be nice to have, but for now I'd settle for Ubuntu recognizing this as a touchpad and allowing users to set desirable touchpad behaviors.
I did a bit of poking with the elantech stuff on the psmouse kernel driver. This leads me to believe that the issue should be reported against kernel, since this is where the problem starts.
When queried, this pad reports a different set of numbers than other Elantech pads, this happens BEFORE the part where the force_elantech parameter comes in action, thus it is useless in this case.
I managed to get the driver to recognize the pad as an Elantech by changing the expected reply; indeed it reports back as "ETPS/2 Elantech Touchpad". However, the data the pad sends is apparently encoded differently, and there's basically no coordinate information. Only thing that works are the buttons, and even then, sporadically; the driver goes out of sync (reported in dmesg). So getting it to work will probably be a bit more involved.
There's already an Ubuntu bug reported for the QX410 https://bugs.launchpad.net/bugs/681904. It's reported against xserver-xorg-video-intel, which I believe is not correct. I'll submit my findings and hopefully this will help get things sorted out.
@Roadmaster, It may be stuck in relative mode, or the protocol may be different again, try fuzzing the register values to get it into absolute mode. If you go to the original bug for this thread here check out comments 23, 24 and 25 for some possibly helpful clues. Feel free to PM me if you want, I was trying to get somewhere with the Samsung, but since I don't have a machine with this touchpad I can't really help much.
This will provide a list of current settings for the touchpad. For instance, you may see TapButton2=3, which means a three-finger strike will trigger mouse button 2. You can change that to two-finger strike triggers button 2 by entering the following code:Code:synclient -l
You can extrapolate this to any of the options present there. Check the man synclient for more info.Code:synclient TapButton2=2
If your touchpad is truly being recognized as a touchpad, then you should have options in system>preferences>mouse>touchpad to disable clicks while typing. If not there, check gconf-editor apps/gnome/peripherals/touchpad as well.
Hope this helps.
Last edited by benbeel; December 23rd, 2010 at 05:00 AM.
Thanks for the pointers on the QX410's touchpad. I have been poking around the reg_10 register but have been unable to get anything from the pad; the elantech.txt doc states that in v2 hardware, bit 1 of reg_1 controls drag & drop, but nothing changes if I touch this bit. if I do tap-hold-drag, the pad still reports button 1 is pressed during the drag.
I modified the driver so that it reports to syslog the decoded absolute coordinates, number of fingers, and status of the buttons. What I see is packets being read from the pad, but coordinates get decoded as 0,0, and fingers is always 0. If I use two fingers, the pad doesn't even generate packets. Only the buttons work as expected, with a tap on the pad doubling as the left button.
I'm also getting a lot of these loss-of-sync events, every time this happens the driver gets reinitialized:
Is there a way to see how the Windows driver initializes and configures the pad? that might help go in the right direction.Code:Dec 23 20:30:17 snowflake kernel: [ 6406.333118] psmouse.c: Touchpad at isa0060/serio1/input0 lost synchronization, throwing 3 bytes away. Dec 23 20:30:18 snowflake kernel: [ 6406.839460] psmouse.c: resync failed, issuing reconnect request Dec 23 20:30:18 snowflake kernel: [ 6407.155017] elantech: retrying ps2 command 0xe6 (2). Dec 23 20:30:19 snowflake kernel: [ 6407.670745] elantech: retrying ps2 command 0xf8 (2). Dec 23 20:30:19 snowflake kernel: [ 6408.375595] elantech: retrying ps2 command 0xf8 (1).
I'll keep exploring and report back.