Bug reported by "TreeBoy" at the URL above 3 weeks ago, same diagnostic observed by another user. Config and log available at the URL above. He says: When running X everything is fine until I touch the touchscreen when X dies. The system is unusable as the display does not refresh and all I see is the non-updating X server window, even though all X processes are no longer running. If I connect from another PC, I get an error along the lines of "Undefined function IsCorePointer" jcristau told him to try the following patch to fix this error: http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=10;filename=fpit-xserver-1.4.diff;att=1;bug=457722 It fixed the crash but the touchscreen still does not work: "Unfortunately, the touchscreen no longer does anything. If I cat /dev/ttyS1 I do see data, but the mouse pointer doesn't respond at all." He also tried setting is_core_pointer to 0, but it didn't help.
I think the driver needs work with the conv_x/conv_y until the scaling is back in the server. ATM it wouldn't work too well. See #10324 Aside from that, I think the call to xf86XInputSetScreen can be removed. Can you give those two a try and see if it changes anything. I don't see anything wrong in the driver other than the scaling thing.
ping?
The submitter replied yesterday on the Debian BTS, here's what he says: I have now tried commenting out the "xf86XInputSetScreen" call which I found on line 259. I tried it with the unadulterated Debian source which caused X to hang: I could only restart the machine by holding down the power button. I then tried it both with the "is_core_pointer=0" and with "is_core_pointer=1". They both produced the same result which is the last result I described: no pointer movement but output appearing from "cat /dev/ttyS1". I'm afraid that I do not understand what "try those two" mean. I presume that the second item is related to freedesktop bug 10324, but I could not identify the test. Many thanks for all your help. (Sorry about the delay in responding, I have been forced to live outside of my home for the last two weeks due to leaking from an upstairs neighbour. This shall mean that I shall also be away from my home between 2nd Feb and 23 Feb while the serious work is done on repairs.)
Created attachment 14026 [details] [review] random patches trying to make fpit work again Please give the above patches a try. I don't know if they work, but they may just do. If the pointer still doesn't move, put an ErrorF("....") in before posting events (look for xf86PostMotionEvent) and print the conv_x/conv_y coordinates.
Hi. I'm the original reporter on Debian. I have tried the patch you have supplied and I get the same problem as last described: no mouse pointer movement, but I can see output going through /dev/ttyS1 I noticed while compiling that there were a couple of problems: dpkg-shlibdeps: warning: symbol xf86PostButtonEvent used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol xf86Msg used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol RemoveEnabledDevice used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol xf86RemoveEnabledDevice used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol InitValuatorClassDeviceStruct used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol xf86SetIntOption used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol Xfree used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol xf86XInputSetScreen used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol InitFocusClassDeviceStruct used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: symbol xf86CollectInputOptions used by debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in none of the libraries. dpkg-shlibdeps: warning: 25 other similar warnings have been skipped (use -v to see them all). dh_gencontrol dpkg-gencontrol: warning: unknown substitution variable ${misc:Depends} dh_md5sums dh_builddeb dpkg-deb: building package `xserver-xorg-input-fpit' in `../xserver-xorg-input-fpit_1.1.0-3+3_i386.deb'. dpkg-genchanges >../xserver-xorg-input-fpit_1.1.0-3+3_i386.changes dpkg-genchanges: not including original source code in upload dpkg-buildpackage: binary and diff upload (original source NOT included) **END I don't know if this is relevant, but I thought I would add it. Many thanks for your assistance. TreeBoy
Created attachment 14040 [details] the source I actually ran 2008/01/30 This is the file that I ran. The diff that you provided did not match my source: I do not know why they are so different, but I hope that this helps.
(In reply to comment #5) > Hi. > > I'm the original reporter on Debian. > > I have tried the patch you have supplied and I get the same problem as last > described: no mouse pointer movement, but I can see output going through > /dev/ttyS1 > > I noticed while compiling that there were a couple of problems: > > dpkg-shlibdeps: warning: symbol xf86PostButtonEvent used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. [...] this is a setup issue. all these functions are defined in the xserver. maybe you're lacking xserver-dev or something like that? but they're not really a problem. (In reply to comment #6) > Created an attachment (id=14040) [details] > the source I actually ran 2008/01/30 > > This is the file that I ran. > > The diff that you provided did not match my source: I do not know why they are > so different, but I hope that this helps. the one I posted is the one for git master. Please get the driver from git://anongit.freedesktop.org/git/xorg/driver/xf86-input-fpit, otherwise it'll make debugging even harder.
(In reply to comment #5) > I noticed while compiling that there were a couple of problems: > > dpkg-shlibdeps: warning: symbol xf86PostButtonEvent used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol xf86Msg used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol RemoveEnabledDevice used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol xf86RemoveEnabledDevice used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol InitValuatorClassDeviceStruct used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol xf86SetIntOption used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol Xfree used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol xf86XInputSetScreen used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol InitFocusClassDeviceStruct used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. > dpkg-shlibdeps: warning: symbol xf86CollectInputOptions used by > debian/xserver-xorg-input-fpit/usr/lib/xorg/modules/input/fpit_drv.so found in > none of the libraries. ignore this, it's a bug in dpkg-shlibdeps imo.
Hello. Sorry about the delay, but I've only just got back home (after having to evacuate for building repairs) and got the machine out of storage. I am now available every evening after work and so you should find me much more responsive. Anyway: I got the package from the git URI that you provided, but I found that the patch that you had provided did not exactly fit with the code that I got (version 1.53 for your patch, but 1.6 i git). I attach the source that I have been using. I got the same results as before: X runs and the driver is loaded, but no movement happens although stuff appears through /dev/ttyS1. When I added the debug messages, I find that both conv_x and conv_y are always zero. Hope this helps. Regards
Created attachment 14864 [details] the source I ran 2008/03/05
Created attachment 15749 [details] [review] Patch for current git tree The same patches, should apply to the current git tree. The coordinate conversion has been merged into the server, so this isn't needed anymore. However, if you don't have a recent server from git, your movements may be wrong.
Hi, I am really trying to fix this problem and really appreciate your work so please keep it up and I will test as I can. If I use the current GIT version of the module, my X doesn't crash but the pointer also doesn't move. As the other poster mentioned catting gives info. If I apply your most recent patch, it crashes my X. I heard there was a way to compile with a debug switch but I am pretty unexperienced in the area and can basically only apply patches and compile, but want to help solve this issue. Let me know what else I can do! Thanks!
Do I need to compile the whole X server from GIT? This is what my log says about my current version. I'm on Ubuntu Hardy on a Gateway CX2619. X.Org X Server 1.4.0.90 Release Date: 5 September 2007 X Protocol Version 11, Revision 0 Build Operating System: Linux Ubuntu (xorg-server 2:1.4.1~git20080131-1ubuntu9) Current Operating System: Linux MobileStudio 2.6.24-16-rt #1 SMP PREEMPT RT Thu Apr 10 15:15:40 UTC 2008 i686 Build Date: 15 April 2008 05:26:17PM
I can now confirm that the current GIT version works. In xorg.conf, the AlwaysCore entry should not exist, and SendCoreEvents should be set to "true". Also no BaudRate should be set (at least for me). The only thing that is not working for me is right-clicking. I added "xsetpointer TOUCHSCREEN xmodmap -e "pointer = 1 3 2"" to .xinitrc in my home directory. When I do "xsetpointer TOUCHSCREEN" in the console, it says "Extended device TOUCHSCREEN not found" Otherwise, everything is working. Any guidance on the right-click issue?
(In reply to comment #15) > I can now confirm that the current GIT version works. In xorg.conf, the > AlwaysCore entry should not exist, and SendCoreEvents should be set to "true". this has been fixed in the server, was cherry-picked yesterday by jcristau. > The only thing that is not working for me is right-clicking. I added > "xsetpointer TOUCHSCREEN > xmodmap -e "pointer = 1 3 2"" > to .xinitrc in my home directory. When I do "xsetpointer TOUCHSCREEN" in the > console, it says "Extended device TOUCHSCREEN not found" try xsetpointer -l and make sure if TOUCHSCREEN is listed in your devices.
(In reply to comment #16) > (In reply to comment #15) > > I can now confirm that the current GIT version works. In xorg.conf, the > > AlwaysCore entry should not exist, and SendCoreEvents should be set to "true". > > this has been fixed in the server, was cherry-picked yesterday by jcristau. > Meaning that once we get the new version of the server, this config will still work, yes? It will just be less picky? > > The only thing that is not working for me is right-clicking. I added > > "xsetpointer TOUCHSCREEN > > xmodmap -e "pointer = 1 3 2"" > > to .xinitrc in my home directory. When I do "xsetpointer TOUCHSCREEN" in the > > console, it says "Extended device TOUCHSCREEN not found" > > try xsetpointer -l and make sure if TOUCHSCREEN is listed in your devices. > jon@MobileStudio:~$ xsetpointer -l 0: "Virtual core keyboard" [XKeyboard] 1: "Virtual core pointer" [XPointer] 2: "Generic Keyboard" [XExtensionKeyboard] 3: "<default pointer>" [XExtensionPointer] 4: "TOUCHSCREEN" [XExtensionPointer] 5: "Synaptics Touchpad" [XExtensionPointer] jon@MobileStudio:~$ xsetpointer TOUCHSCREEN Extended device TOUCHSCREEN not found jon@MobileStudio:~$ xsetpointer "TOUCHSCREEN" Extended device TOUCHSCREEN not found .......
(In reply to comment #17) > (In reply to comment #16) > > (In reply to comment #15) > > > I can now confirm that the current GIT version works. In xorg.conf, the > > > AlwaysCore entry should not exist, and SendCoreEvents should be set to "true". > > > > this has been fixed in the server, was cherry-picked yesterday by jcristau. > > > Meaning that once we get the new version of the server, this config will still > work, yes? It will just be less picky? yes, the old handling was simply incorrect. > jon@MobileStudio:~$ xsetpointer -l > 0: "Virtual core keyboard" [XKeyboard] > 1: "Virtual core pointer" [XPointer] > 2: "Generic Keyboard" [XExtensionKeyboard] > 3: "<default pointer>" [XExtensionPointer] > 4: "TOUCHSCREEN" [XExtensionPointer] > 5: "Synaptics Touchpad" [XExtensionPointer] > jon@MobileStudio:~$ xsetpointer TOUCHSCREEN > Extended device TOUCHSCREEN not found > jon@MobileStudio:~$ xsetpointer "TOUCHSCREEN" > Extended device TOUCHSCREEN not found this is a bug in xsetpointer. I'm fixing it right now, but with server 1.4 and above xsetpointer is deprecated to a point that it can't do anything but listing the devices. In your case, you don't need this call.
> > > jon@MobileStudio:~$ xsetpointer -l > > 0: "Virtual core keyboard" [XKeyboard] > > 1: "Virtual core pointer" [XPointer] > > 2: "Generic Keyboard" [XExtensionKeyboard] > > 3: "<default pointer>" [XExtensionPointer] > > 4: "TOUCHSCREEN" [XExtensionPointer] > > 5: "Synaptics Touchpad" [XExtensionPointer] > > jon@MobileStudio:~$ xsetpointer TOUCHSCREEN > > Extended device TOUCHSCREEN not found > > jon@MobileStudio:~$ xsetpointer "TOUCHSCREEN" > > Extended device TOUCHSCREEN not found > > this is a bug in xsetpointer. I'm fixing it right now, but with server 1.4 and > above xsetpointer is deprecated to a point that it can't do anything but > listing the devices. In your case, you don't need this call. > Well, if I take out that line of my .xinitrc, the button on the pen still registers as a middle click, not a right click.
(In reply to comment #19) > Well, if I take out that line of my .xinitrc, the button on the pen still > registers as a middle click, not a right click. All this line did was to set the TOUCHSCREEN to be the core pointer. This isn't necessary anymore (and not possible anyway). actually - you may want to skip the xmodmap line. It's mapping button 3 to 2, and looking at the fpit code this may be mapping right button to middle button.
(In reply to comment #20) > (In reply to comment #19) > > Well, if I take out that line of my .xinitrc, the button on the pen still > > registers as a middle click, not a right click. > > All this line did was to set the TOUCHSCREEN to be the core pointer. This isn't > necessary anymore (and not possible anyway). > > actually - you may want to skip the xmodmap line. It's mapping button 3 to 2, > and looking at the fpit code this may be mapping right button to middle button. > The pen button by default is a middle click. I want to change this to make it right-click. With or without the xmodmap line the button on the pen performs a middle-click.
(In reply to comment #21) > > actually - you may want to skip the xmodmap line. It's mapping button 3 to 2, > > and looking at the fpit code this may be mapping right button to middle button. > > > > The pen button by default is a middle click. I want to change this to make it > right-click. With or without the xmodmap line the button on the pen performs a > middle-click. > The button on the pen registers in xinput test TOUCHSCREEN as a button 2 press. Running the xmodmap line from a terminal produces the desired effect, except that it does it for my touchpad as well. I only want it to do this for the pen. Perhaps modifying the source is a better alternative to this. But I'm still not sure why it doesn't work from my .xinitrc file.
(In reply to comment #22) > The button on the pen registers in xinput test TOUCHSCREEN as a button 2 press. button 2 is middle click. > Running the xmodmap line from a terminal produces the desired effect, except > that it does it for my touchpad as well. I only want it to do this for the pen. xmodmap modifies the virtual core pointer. the physical devices all feed into this device, thus it applies to all devices. > Perhaps modifying the source is a better alternative to this. But I'm still not > sure why it doesn't work from my .xinitrc file. no. this shouldn't be in the driver. what we need is a modification of xmodmap to do this per device. on that note, xmodmap is fairly straightforward, so you can get cracking right away ;) Pushing the fixes as 2823f7c3f3784a1a398b8d5a9157c2e5263b6285 01a4da993a2ecd0f017baef49b6e57b82c05316d 95676d89c7954cc5550ed386b784b470034320cc Bug marked as fixed.
Thanks for all your help. I achieved the button switch with "xinput set-button-map TOUCHSCREEN 1 3 2" This will do for now.
hi there savvy guys, thanks to the patch provided by Peter Hutterer I was finally able to get my digitizer and the right-clicking issue working on a cx2735m Gateway, nevertheless, I cannot get any clicking whenever I revert the screen to Tablet mode by doing "xrandr -o 2", but it does track the pointer. Any suggestions???... Thanks in advance, you've been extremely helpful
(In reply to comment #25) > hi there savvy guys, thanks to the patch provided by Peter Hutterer I was > finally able to get my digitizer and the right-clicking issue working on a > cx2735m Gateway, nevertheless, I cannot get any clicking whenever I revert the > screen to Tablet mode by doing "xrandr -o 2", but it does track the pointer. > Any suggestions???... good question. please start up xev and see if any clicks show up there. If not, please open a separate bug with the xev output and a xorg log file attached. (lets try to keep bugreports to one problem at a time).
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.