On Sun, Nov 13, 2011 at 10:16 AM, Cedric Sodhi <man...@gmx.net> wrote: > On Sun, Nov 13, 2011 at 09:21:20AM -0600, Chris Bagwell wrote: >> On Sat, Nov 12, 2011 at 9:42 AM, Cedric Sodhi <man...@gmx.net> wrote: >> > Chris, I just compiled from Torvalds' HEAD (adding MULTITOUCH5 as usual) >> > and the clicking still jams. >> > >> >> Hmm, I just tried that version and its working with my touchscreen (I >> got it back). I forget, what X input driver are you using and what >> version of it? >> > > Evdev 2.6.0 > >> I'm using a locally patched version of xf86-input-wacom so maybe thats >> why its working for me. > > Are you implying that there is a chance that the wacom driver works for > the eGalax?
Yes, it is technically possible to use xf86-input-wacom with a multitouch touchscreen. You have to edit a config file since by default xf86-input-evdev handles them. There are some rough edges though if you do it. > >> >> Can you use evtest and monitor the events coming out when you release >> your touch? This will help isolate if its kernel driver issues. Upon >> release, do you see Touch change from 1 to 0 or is there no event? If >> no event, do you see Touch go from 1 to 0 and then immediately back to >> 1 once you touch screen again? > > It looks like the jamming has stopped. Maybe it was something else last > time I checked. However, no clicks don't work at all. Xinput test and > evtest don't show any clicking, if I'm not mistaken. Log of evtest > attached. For xinput you just have to believe me since I don't know how > to capture its output. Was the attached evtest log a complete output of you touching the screen and releasing? The reason I ask is because there were no Touch reports in the output. Just X/Y movement. If you did touch and release, it does show its the kernel driver having issues. This would explain your stuck ON or stuck OFF button press. It would be stuck ON or OFF based on if you first touched the screen before X is started or restarted (stuck off) or after X was started up (stuck on). There is a MT_QUIRK_NOT_SEEN_MEANS_UP that you can add to MT_CLS_EGALAX definition to see if that helps. > >> >> I also find using the kernel hid debug features useful in >> understanding whats going on: >> >> sudo cat /sys/kernel/debug/hid/*/events > > I don't get any output on that (I only got two devices in ./hid - > shouldn't there be more? Keyboard, wacom, egalax, usbcamera, etc etc). You'll only get them for devices handled by HID drivers. On my tablet, thats only 1 touchscreen. I was in fact assuming you only have 1 HID device. Can you retry this and do the exact filename? Also, you will need to cat the "rdesc" files in same directory. Probably the above is catching your non-eGalax device and so your not seeing events. The events are now very important to see to understand how its reporting touches. We may even have to go lower and get raw USB packets since the debug output h In my rdesc, you can see Touch is reported by the TipSwitch event: Digitizers.TipSwitch ---> Key.Touch And when I touch the screen, you see this in my events output; which is my I see Touch events: Digitizers.TipSwitch = 1 >> >> This will give you a human readable version of HID packet as they come >> in from hardware. >> >> > Even further: I noticed that there appears to be no precedence setting >> > as for the Touchscreen or the wacom pen. With the integrated Wacom >> > Finger/Pen combination the Wacom succeessfully (at least partly, see >> > that old unresolved bug) handled that internally, but now that these are >> > actually two different devices the touch interferes with the wacom >> > (meaning I can't write with my palm on the screen - not to mention I >> > still cannot use the eGalax at all). >> > >> > Any ideas? >> >> Ping added support to xf86-input-wacom a while back were if it detects >> a pen+touch that are the same device then it will mask touch during >> pen usage. It would be pretty easy to add an interface were user can >> inform that two devices are related when it can't be auto-detected. > > It still sounds like you are planning to use the Wacom Driver for both, > the Wacom and the eGalax. Can you clarify that? Yes, that would be the easiest solution for this hardware combo (smallest amount of missing code to write). The other alternative is to write a daemon similar to synclient. It would monitor for motion on 1 device and when seen it would disable another device. Similar to how synclient can disable a touchpad when your typing. Chris >> >> We can discuss that more once you've got the touch fully working. >> >> In mean time, you can use 'xinput set-prop "name of touchscreen" >> "Device Enabled" 0' to disable the touchscreen manually when your >> going to use stylus. I know; not ideal. > > That's how I've been doing it so far. > > -- > Cedric > ------------------------------------------------------------------------------ RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 _______________________________________________ Linuxwacom-devel mailing list Linuxwacom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel
