Ubuntu on Eee PC T101MT

[Plippo]

I did everything in that post up to the install calibrate step.

A few days ago, I updated the instructions and uploaded a new version of the driver which no longer needs calibration. I recommend simply re-installing the touchscreen support package and the kernel driver, this should solve the problem.

[Plippo]

@tlimsisnw: Oops, I didn't see your comment before that the keyboard doesn't work when twofing is not running. This is strange, so the problem can't only be twofing and grab-and-drag but there must be a different configuration problem with your system.

[riccardo.depalo]

@Plippo: I was watching your script for installing egalax driver and produce a deb file. Maybe it tries to cd into the source and doesn't seem to find it because the file is called just linux-source-2.6.35.tar.bz2 and not 2.6.35-19-generic as my system says when I type uname -r? maybe that's the reason why it doesn't seem to work in Maverick?

[Kevrox]

Hi Plippo
I now have added a twofing launch on the desktop and removed the launcher I had in startup applications, so all I do after a reboot is double click on the icon I made and twofing works .... not the best way but a way non the less, wonder why it does not want to run when set up in startup applications.
cheers and mate thanks for all the hard work you do, all we do is test and moane you do all the fixing.

[gltripp]

@Kev: If this rectangle appears, it means that the taps of both fingers are registered as single clicks, which probably means that the twofing daemon either doesn't run or doesn't work. If you press Alt+F2, enter twofing and press enter and the two-finger click works then, it means that twofing probably crashed or wasn't loaded correctly before.

@tlimsisnw: This is interesting, I don't know how grab-and-drag works, but if it tries to grab the pointer like twofing does, this could be an explanation. At the moment I don't think there is something to do about it other than disabling either grab-and-drag or twofing.

@gltripp: The multitouch kernel module is only the first step. It sends multitouch events to the X server, but currently X doesn't know how to handle them. What is now required is support in the higher layers of the software stack. There are several attempts to provide this. One of the most advanced seems to be the UTouch framework that will be introduced in Ubuntu Maverick, but I haven't tested that yet. Some months ago, I created the twofing daemon as an intermediate step. It reads the multitouch events directly from the Kernel and translates them into mouse and key events. But currently, it only works for the eGalax touchscreen, not your model.

So if you want to use Multitouch right now, you can either try Maverick and check if the UTouch system already works, which I don't know, or if you want to use twofing, I can try to extend it to also support your screen. As the events should have the same structure, this shouldn't be very much work. If you like, you can send me a PM so I can tell you which additional information I need.

Hi Plippo,

i'll try the modify-twofing-daemon-way

But I'll need some help:

1. i have to modify your udev-rule:
BUS=="usb",ACTION=="add",KERNEL=="event*",SYSFS{id Product}=="480d",SYMLINK+="twofingtouch",RUN+="/bin/chmod a+r /dev/twofingtouch"
to match the corresponding DeviceID of my controller:
SUBSYSTEM=="usb",ACTION=="add",KERNEL=="event*",AT TR{idProduct}=="0186",SYMLINK+="twofingtouch",RUN+ ="/bin/chmod a+r /dev/twofingtouch"

(I changed the SYSFS & BUS-attributes because Udev told me they are deprecated)
But: not device /dev/twofingtouch will be created. Whats wrong ?
I got the value "0186" from 'lsusb' and searching in /sys.

2nd Question:
Why does the driver creates 3 event-devices in /dev/input ?

evtest /dev/input/event7
Input driver version is 1.0.0
Input device ID: bus 0x3 vendor 0x486 product 0x186 version 0x100
Input device name: "AsusTek, Inc. MultiTouch(YFO)"
Supported events:
  Event type 0 (Sync)
  Event type 1 (Key)
    Event code 330 (Touch)
  Event type 3 (Absolute)
    Event code 0 (X)
      Value   1877
      Min        0
      Max     3478
    Event code 1 (Y)
      Value   1886
      Min        0
      Max     3478
    Event code 53 (?)
      Value      0
      Min        0
      Max     3478
    Event code 54 (?)
      Value      0
      Min        0
      Max     3478
    Event code 57 (?)
      Value      0
      Min        0
      Max        2
  Event type 4 (Misc)
    Event code 4 (ScanCode)

evtest /dev/input/event8
Input driver version is 1.0.0
Input device ID: bus 0x3 vendor 0x486 product 0x186 version 0x100
Input device name: "AsusTek, Inc. MultiTouch(YFO)"
Supported events:
  Event type 0 (Sync)
  Event type 1 (Key)
    Event code 272 (LeftBtn)
    Event code 273 (RightBtn)
  Event type 3 (Absolute)
    Event code 0 (X)
      Value      0
      Min        0
      Max     3478
    Event code 1 (Y)
      Value      0
      Min        0
      Max     3478
    Event code 53 (?)
      Value      0
      Min        0
      Max     3478
    Event code 54 (?)
      Value      0
      Min        0
      Max     3478
  Event type 4 (Misc)
    Event code 4 (ScanCode)

evtest /dev/input/event9
Input driver version is 1.0.0
Input device ID: bus 0x3 vendor 0x486 product 0x186 version 0x100
Input device name: "AsusTek, Inc. MultiTouch(YFO)"
Supported events:
  Event type 0 (Sync)
  Event type 3 (Absolute)
    Event code 40 (Misc)
      Value      0
      Min        0
      Max      255

What are about these?
Only event7 is generating any events, when I'm scratching on the display....

[Plippo]

@riccardo: This should not be the problem, the script uses sed to cut the version number. There are two variables in the script, VERSION which contains the full version number and VER which contains the abbreviated one. One thing you could do is replace all occurrences of $VER (but not of $VERSION, so you have to do it by hand) in the script by 2.6.35, maybe the variable is causing the problem as this is the only thing I changed.

@gltripp: The three devices might be there because you are maybe still using the old QUIRKS option. If you remove it, there should only be one device. Maybe this is also what confuses udev.

Altering the udev rule is exactly what you have to do. Are you using Lucid or Maverick? In Lucid, it still works with the deprecated keywords. At first glance, everything looks okay, but I know that udev rules are unfortunately really hard to debug. The only thing that looks wrong are the spaces, but I guess they were added when you posted here?

EDIT: You can also try the command udevadm info --query=all --name=/dev/input/event7 --attribute-walk which lists all attributes for the device. The output suggests that you should use ATTRS instead of ATTR in the rule, but this is just a guess.

[gltripp]

@gltripp: The three devices might be there because you are maybe still using the old QUIRKS option. If you remove it, there should only be one device. Maybe this is also what confuses udev.

*hm* Mea culpa!
This was my fault!

Altering the udev rule is exactly what you have to do. Are you using Lucid or Maverick? In Lucid, it still works with the deprecated keywords. At first glance, everything looks okay, but I know that udev rules are unfortunately really hard to debug. The only thing that looks wrong are the spaces, but I guess they were added when you posted here?

EDIT: You can also try the command udevadm info --query=all --name=/dev/input/event7 --attribute-walk which lists all attributes for the device. The output suggests that you should use ATTRS instead of ATTR in the rule, but this is just a guess.

No, sorry - this didn't worked too ;-(

SUBSYSTEM=="usb",ACTION=="add",KERNEL=="event*",AT TRS{idVendor}=="0486",ATTRS{idProduct}=="0186",SYM LINK+="twofingtouch",RUN+="/bin/chmod a+r /dev/twofingtouch"

damn it

[fastler]

A few days ago, I updated the instructions and uploaded a new version of the driver which no longer needs calibration. I recommend simply re-installing the touchscreen support package and the kernel driver, this should solve the problem.

Nope, I installed the package and drivers yesterday. I reinstalled them again today, but nothing has changed. The system recognizes that I have touched the screen, but it thinks that I have touched the very bottom right corner of the screen. This is a big pity, because everything else is working a charm, just the touchscreen isn't. Sorry to be a hassle.

[Plippo]

damn it

Don't worry, we'll beat the udev monster. And if not, there is still the possibility to create the link manually in the boot scripts. But of course udev would be the better solution. Could you maybe post the output of

udevadm info --query=all --name=/dev/input/event7 --attribute-walk

(replace the event7 if the device number has changed after removing the QUIRKS option)

@fastler: This is strange. What happens when you enter

xinput --list

and

xinput --list-props 10 && xinput --list --long 10

in the terminal?

[riccardo.depalo]

@Plippo: no, it didn't work. There may be be something wrong in linux-source. But I controlled: headers, image and source are all 2.6.35-20.29 This is my output:

./build.sh
BUILDING KERNEL MODULE...
Cleaning up...
rm: cannot remove `linux-source-2.6.35.tar': No such file or directory
Unpacking linux kernel source...
bunzip2: Can't open input file linux-source-2.6.35.tar.bz2: No such file or directory.
tar: linux-source-2.6.35.tar: Cannot open: No such file or directory
tar: Error is not recoverable: exiting now
Copying Module.symvers from headers...
cd: 93: can't cd to linux-source-2.6.35
Configuring...
make: *** No rule to make target `oldconfig'.  Stop.
Building scripts...
make: *** No targets.  Stop.
grep: include/linux/hid.h: No such file or directory
sed: can't read include/linux/hid.h: No such file or directory
Modified hid.h.
cd: 93: can't cd to drivers/hid
grep: hid-core.c: No such file or directory
sed: can't read hid-core.c: No such file or directory
Modified hid-core.c.
grep: usbhid/hid-quirks.c: No such file or directory
hid-quirks.c is already fine.
hid-ids.h is already fine.
Makefile is already fine.
Building modules...
make: Entering directory `/'
make: *** No rule to make target `modules'.  Stop.
make: Leaving directory `/'
Copying...
cp: cannot stat `hid.ko': No such file or directory
cp: cannot stat `hid-egalax.ko': No such file or directory
cp: cannot stat `usbhid/usbhid.ko': No such file or directory
Should have built. Ready to build deb.
BUILDING DEB PACKAGE...
dpkg-deb: generazione del pacchetto "egalax-multitouch-driver-2.6.35-19-generic-i386" in "deb.deb".
SHOULD BE FINISHED.