Touch Screen Issue

[jwelch1324]

Hey everyone,

I have an unusual issue with two touchscreen based MD's. One of them uses a penmount panel (with latest 8.10 ubuntu drivers) and the other is a TSHARC panel (again with latest drivers)

The issue is the following, If I boot the MD's into the orbiter (either with reboot, or by a a killall xinit && /usr/pluto/bin/Start_X.sh command) the touch screens will not function properly unless I touch the panel the moment X starts before the oribiter starts to load. As long as I register a click before the orbiter appears they work fine... but I haven't been able to figure out why...

Anyone have any thoughts and or ideas on a workaround (maybe script that simulates a click event before the orbiter loads??)

Thanks
~jw

[l3mce]

I expect some sort of modprobe would handle it.

Not sure what specifically to probe or where to probe it.

[jwelch1324]

Solved:

Problem was the savage driver on the penmount display, I changed in xorg.conf under "Device" section

Code: [Select]


Section "Device"
        Driver          "savage"
# ...
        Option          "SWcursor"      "on"                    # [<bool>]
        Option          "HWCursor"      "off"
# ...
EndSection

Forcing it to use the SW cursor fixes the problem with the penmount panel.... EDIT: No it doesn't... it just makes sure the cursor is under my finger when I touch the screen, the issue of having to touch the panel before the orbiter loads is still there...

On the TSHARC panel, the issue was actually that modprobe (thx l3mce for the advice to look into this   ), or more specifically KMS was detecting a monitor port that didn't exist, and hence cycling a connect/disconnect from a monitor that didn't exist. (The chipset supports the extra VGA output, but the panel computer doesn't have an actual VGA out port on it, this is actually a bug in the intel driver used in ubuntu 810). But either way to circumvent the issue I added the following to xorg.conf

Code: [Select]


Section "Monitor"
        Identifier      "foo"
        Option  "Ignore" "true"
EndSection

Section "Device"
#...

        Driver          "intel"
#...
        Option "monitor-VGA" "foo"
        Option "monitor-LVDS" "Monitor0"
#...
EndSection

The "Ignore" directive tells xorg not to load the monitor into xrandr, and hence it won't attempt to detect any dynamic connect/disconnect events on that monitor, which is assigned to the "monitor-VGA" output of the chip.

Now xrandr only outputs the LVDS monitor as being present... SUCCESS!!

Thanks,
~jw

[JaseP]

Jwelch:

What kind of touchscreen controller does it have,... an internal PS/2 serial or USB???

Also, could you post your xorg.conf???

With serial touchscreen controllers in some newer versions of Linux, there is a problem with the /udev system grabbing the serial port for a mouse pointer. I'm not so convinced that that is the problem with a system based on Ubuntu 8.10. Another issue is that some of the touch drivers want to be primary. that would be set in the mouse pointer area of your xorg.conf.

As far as doing a modprobe or insmod, etc., is that most of the driver installation scripts will do that for you...

Please post the text file output of the command:

Code: [Select]

lsmod > installed_modules.txt

Thanks,...

[JaseP]

Solved:

...

The "Ignore" directive tells xorg not to load the monitor into xrandr, and hence it won't attempt to detect any dynamic connect/disconnect events on that monitor, which is assigned to the "monitor-VGA" output of the chip.

Now xrandr only outputs the LVDS monitor as being present... SUCCESS!!

Thanks,
~jw

I was a little late with my help...
Glad you fixed it.

[jwelch1324]

JaseP,

Actually the problem wasn't completely solved as I previously thought, it turns out on reboot, the penmount device still wasn't working without first touching the screen prior to the orbiter loading.

So to answer some of your questions,

The interface is serial, on /dev/ttyS2

the output from lsmod doesn't have any kernel modules for the penmount device loaded (it is only loaded as module by Xorg), but just for the sake of curiosity in case it gives you some insight as to the source of the issue here is the lsmod output

Code: [Select]

root@moon158:~# lsmod
Module                  Size  Used by
sco                    18308  2
bridge                 56980  0
stp                    10628  1 bridge
bnep                   20352  2
rfcomm                 44432  0
l2cap                  30464  6 bnep,rfcomm
bluetooth              62180  6 sco,bnep,rfcomm,l2cap
nfsd                  231024  13
auth_rpcgss            42528  1 nfsd
exportfs               12544  1 nfsd
ppdev                  15748  0
lp                     17156  0
autofs4                27780  1
ipv6                  264356  10
longhaul               17896  0
cpufreq_stats          13188  0
cpufreq_powersave       9856  0
cpufreq_ondemand       14988  1
freq_table             12672  3 longhaul,cpufreq_stats,cpufreq_ondemand
cpufreq_conservative    14600  0
cpufreq_userspace      11396  0
container              11520  0
ac                     12292  0
video                  25488  0
output                 11008  1 video
pci_slot               12680  0
sbs                    19464  0
sbshc                  13440  1 sbs
wmi                    14504  0
battery                18436  0
joydev                 18368  0
evdev                  17696  4
button                 14224  0
i2c_viapro             15764  0
i2c_core               31892  1 i2c_viapro
pcspkr                 10624  0
parport_pc             39332  1
shpchp                 38036  0
via_agp                16256  1
pci_hotplug            34976  1 shpchp
parport                42604  3 ppdev,lp,parport_pc
via686a                21004  0
agpgart                42184  1 via_agp
af_packet              25856  0
nfs                   269416  2
lockd                  71976  2 nfsd,nfs
nfs_acl                11264  2 nfsd,nfs
sunrpc                196960  23 nfsd,auth_rpcgss,nfs,lockd,nfs_acl
usbhid                 35712  0
hid                    50560  1 usbhid
sd_mod                 42392  0
crc_t10dif              9984  1 sd_mod
sg                     39732  0
ata_generic            12932  0
pata_acpi              12160  0
pata_via               16260  0
libata                178336  3 ata_generic,pata_acpi,pata_via
uhci_hcd               30864  0
8139too                31616  0
floppy                 64644  0
scsi_mod              155468  3 sd_mod,sg,libata
8139cp                 27520  0
mii                    13440  2 8139too,8139cp
usbcore               149616  3 usbhid,uhci_hcd
dock                   16656  1 libata
raid10                 30464  0
raid456               135184  0
async_xor              11520  1 raid456
async_memcpy           10112  1 raid456
async_tx               15312  3 raid456,async_xor,async_memcpy
xor                    23688  2 raid456,async_xor
raid1                  30080  0
raid0                  15488  0
multipath              15104  0
linear                 13440  0
md_mod                 94108  6 raid10,raid456,raid1,raid0,multipath,linear
thermal                23708  0
processor              42156  2 thermal
fan                    12548  0
nbd                    19088  0
fuse                   60956  1
fbcon                  47648  0
tileblit               10880  1 fbcon
font                   16512  1 fbcon
bitblit                13824  1 fbcon
softcursor              9984  1 bitblit

My xorg.conf is

Code: [Select]

Section "ServerLayout"
Identifier "XFree86 Configured"
Screen   0 "Screen0" 0 0
# InputDevice "Mouse0" "CorePointer"
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Penmount" "CorePointer"
EndSection

Section "Files"
# RgbPath "/usr/X11R6/lib/X11/rgb"
ModulePath "/usr/lib/xorg/extra-modules,/usr/lib/xorg/modules"
FontPath "/usr/share/fonts/X11"
FontPath "/usr/X11R6/lib/X11/fonts/misc/"
FontPath "/usr/X11R6/lib/X11/fonts/Speedo/"
FontPath "/usr/X11R6/lib/X11/fonts/Type1/"
FontPath "/usr/X11R6/lib/X11/fonts/CID/"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi/"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi/"
EndSection

Section "Module"
Load "dbe"
Load "extmod"
Load "record"
Load "xtrap"
Load "speedo"
# Load "type1"
Load "glx"
Load "int10"
Load "v4l"
Load "dri"
# Load "penmount"
EndSection

#Section "InputDevice"
# Identifier "Mouse0"
# Driver "void"
# Option "Device" "/dev/input/mice"
#EndSection

Section "InputDevice"
Identifier "Keyboard0"
Driver "keyboard"
Option "XkbLayout" "us"
Option "XkbVariant" "basic"
Option "XkbModel" "pc105"
EndSection

Section "InputDevice"
 Identifier  "Penmount"
  Driver    "penmount"
  Option  "Device"           "/dev/ttyS2"
  Option  "Protocol"         "PM9000"
  Option  "BaudRate"         "19200"
  Option  "ADBit"            "10"
  Option  "ConfigFile"       "/etc/penmount.dat"
  Option  "Button2"          "3"  # 1=left, 2=middle, 3=right
  Option  "PenDownMode"      "0"  # 0=stream mode, 1=point mode
  Option  "HoldTime"         "2000"  # how long ms to launch the 2nd button
  Option  "ScreenScale"      "0"  # screen scale enable/disable
  Option  "LockWindowRange"  "32"  # range for press and hold
  Option  "CalibHoldTime"    "1000"  # hold time for calibration
  Option  "XMinOffset"       "10"  # edge compensation
  Option  "XMaxOffset"       "10"
  Option  "YMinOffset"       "10"
  Option  "YMaxOffset"       "10"
  Option  "StdXMin"          "30"  # standard calibration
  Option  "StdXMax"          "990"
  Option  "StdYMin"          "30"
  Option  "StdYMax"          "990"
  Option  "Beep"             "1"  # 0=off, 1=down, 2=up, 3=down+up
  Option  "PressVol"         "100"  # 0 = slience
  Option  "PressPitch"       "880"  # freq. (Hz)
  Option  "PressDur"         "15"  # duration (1~255)
  Option  "ReleaseVol"       "0"  # 0 = slience
  Option  "ReleasePitch"     "1200"  # freq. (Hz)
  Option  "ReleaseDur"       "10"  # duration (1~255)
  Option  "AveragePoint"     "0"  # average point (0,4,8)
  Option  "Rotation"         "1"  # rotation (1,2,4,8)
  Option  "RandR"            "0"  # RandR enable/disable)
  Option  "DebugLevel"       "0"  # debug
EndSection


Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown Monitor Vendor"
ModelName "Unknown Monitor Model"
HorizSync 20-500
VertRefresh 59-61
EndSection

Section "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz"
### [arg]: arg optional
#Option "NoAccel" # [<bool>]
Option "SWcursor" "on" # [<bool>]
Option "HWCursor" "off"
#Option "ColorKey" # <i>
#Option "CacheLines" # <i>
#Option "Dac6Bit" # [<bool>]
#Option "DRI" # [<bool>]
#Option "NoDDC" # [<bool>]
#Option "ShowCache" # [<bool>]
#Option "XvMCSurfaces" # <i>
Identifier "Card0"
Driver "savage"
VendorName "Unknown Vendor Name"
BoardName "Unknown Board Name"
Option "LinearAlloc" "16384"
#BusID "PCI:0:2:0"
# Option "NoDDC" "true"
Option "XvmcUsesTextures" "true"
Option "renderAccel" "true"
Option "NoDDCValue"
Option "UseEDID" "true"
Option "ExactModeTimingsDVI" "true"
Option "NoLogo" "true"
Option "NoBandWidthTest" "true"
Option "ModeValidation" "NoDFPNativeResolutionCheck, NoEdidMaxPClkCheck, NoMaxPClkCheck, AllowInterlacedModes, AllowNon60HzDFPModes, NoEdidModes"
Option "DynamicTwinView" "false"
Option "UseEvents" "true"
Option "ConnectedMonitor" "CRT"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Modes "640x480"
Virtual 640 480
EndSubSection
Option "XvmcUsesTextures" "true"
Option "renderAccel" "true"
Option "NoDDCValue"
Option "UseEDID" "true"
Option "ExactModeTimingsDVI" "true"
Option "NoLogo" "true"
Option "NoBandWidthTest" "true"
Option "ModeValidation" "NoDFPNativeResolutionCheck, NoEdidMaxPClkCheck, NoMaxPClkCheck, AllowInterlacedModes, AllowNon60HzDFPModes, NoEdidModes"
Option "DynamicTwinView" "false"
Option "UseEvents" "true"
Option "TVStandard" "640x480 (4:3)"
EndSection

Section "ServerFlags"
Option "AllowEmptyInput" "true"
Option "AllowMouseOpenFail" "true"
Option "BlankTime" "0"
Option "StandbyTime" "0"
Option "SuspendTime" "0"
Option "OffTime" "0"
Option "AIGLX" "Off"
Option "AutoAddDevices" "False"
EndSection
Section "Extensions"
Option "RENDER" "true"
EndSection

I added the lines "AllowEmptyInput" and "AutoAddDevices" since in my Xorg log it was attempting to add a default pointer even though I had one setup. I thought this might be the issue... sadly it didn't help.

What I do notice that is odd (or maybe not?) is in the Xorg log it looks like it is adding the penmount device twice

Code: [Select]

(II) penmount: PreInit
(**) Option "CorePointer"
(**) Penmount: always reports core events
(**) Option "XMinOffset" "10"
(**) Option "XMaxOffset" "10"
(**) Option "YMinOffset" "10"
(**) Option "YMaxOffset" "10"
(**) Option "StdXMin" "30"
(**) Option "StdXMax" "990"
(**) Option "StdYMin" "30"
(**) Option "StdYMax" "990"
(**) Option "ScreenScale" "0"
(**) Option "Rotation" "1"
(**) Option "RandR" "0"
(**) Option "AveragePoint" "0"
(**) Option "Beep" "1"
(**) Option "PressVol" "100"
(**) Option "PressPitch" "880"
(**) Option "PressDur" "15"
(**) Option "ReleaseVol" "0"
(**) Option "ReleasePitch" "1200"
(**) Option "ReleaseDur" "10"
(**) Option "HoldTime" "2000"
(**) Option "CalibHoldTime" "1000"
(**) Option "LockWindowRange" "32"
(**) Option "Button2" "3"
(**) Option "PenDownMode" "0"
(**) Option "ADBit" "10"
(**) Option "DebugLevel" "0"
(**) Option "ConfigFile" "/etc/penmount.dat"
(II) penmount chip : 9000
(**) Option "Device" "/dev/ttyS2"
(**) Option "BaudRate" "19200"
(**) Option "StopBits" "1"
(**) Option "DataBits" "8"
(**) Option "Parity" "None"
(**) Option "Vmin" "3"
(**) Option "Vtime" "1"
(**) Option "FlowControl" "None"

This one contains all the proper options as set in the xorg.conf file. However just a few lines below that

Code: [Select]

(II) evaluating device (Penmount)
(II) XINPUT: Adding extended input device "Penmount" (type: TOUCHSCREEN)
(II) penmount: min_x=0, max_x=2047, min_y=0, max_y=2047
(**) Option "Device" "/dev/ttyS2"
(**) Option "BaudRate" "19200"
(**) Option "StopBits" "1"
(**) Option "DataBits" "8"
(**) Option "Parity" "None"
(**) Option "Vmin" "3"
(**) Option "Vtime" "1"
(**) Option "FlowControl" "None"

Which doesn't contain all the necessary options (like the calibration file for example). I am not sure if this is normal and XINPUT is just evaluating the device after the server has loaded all the config data, or if it is actually trying to add it as a second device. The fact that I have AutoAddDevices turned off suggests that it is normal behavior but nonetheless the issue is still there. I have to touch the screen before the orbiter progressbar finishes filling up for it to work properly.

Another thing to note is that once I get the touchscreen working properly, I can run killall xinit && /usr/pluto/bin/Start_X.sh and not have to do anything and the touchscreen works fine. It's only if I reboot that the problem shows up again.

Any thoughts are much appreciated...

[JaseP]

Can you post to me the contents of the ReadMe file and the install script that came with the penmount driver???

[JaseP]

Note:

I found this,... use at your own risk;

https://launchpad.net/ubuntu/intrepid/i386/xserver-xorg-input-penmount/1:1.3.0-1build1

I'm not sure if the build is compatible or provides anything better than what you can get in the Intrepid repositories.

There's also Penmount's download page, which appears to be more up-to-date drivers,...

http://www.penmount.com/Download/Driver/PenMount/

Again, I'm not sure if you want to try them, or just plow through the documentation for clues on how to fix your problem.

[jwelch1324]

I am using the Driver for Ubuntu 8.10 from the penmount site. The xorg-input package I tried as well it doesn't work period (i am not sure if there is a discrepancy in the xorg version used by LMCE and the one in the official 8.10 release. Though I can't imagine why it would be)

I will look through all the documentation and see if i can find something about it... I really think it is coming down the SDL renderer used by the orbiter. Because if I reboot the MD and let it load without touching the screen the mouse gets stuck on the left side of the screen. If I then kill xserv and restart it using the pluto script (again without touching the screen) it still gets stuck.

However, if I kill the xserv and then do a normal startx to boot into KDE it works fine.

Furthermore, if I start the MD, touch the screen before the orbiter finishes loading, then kill the xserv and restart using the pluto script it still works fine. However if I then do a quick reload on the router and let it load without touching the screen, it goes back to having the cursor stuck to the left side of the screen.

Most odd...

I was thinking of looking at the padorbiter source to see how it is configured since it uses a penmount screen as well, but can't find it anywhere in the SVN (perhaps it isn't public?).

[JaseP]

Do me a favor and do an lsmod 1st when it's working and then when it isn't...

I'm thinking that you will eventually have to force the penmount driver to load as part of the init script... but I would really like to see the difference between the running modules in the two instances...

[jwelch1324]

Attached is the lsmod output when it is working and when it is not working... sadly they appear to be the same (#diff lsnotworking.txt lsworking.txt gives no changes). The penmount module is not loaded as a kernel module, the driver package from the penmount site doesn't include a kernel module.

I did find a way around the router reload issue, though its a nasty hack. I essentially cron the following script to run every minute to see if there was a request from the router to reload.

Code: [Select]

#!/bin/bash

logfile=/var/log/pluto/LaunchManager.log
timefile=/var/run/reqres
pidfile=/var/run/plutoX0.pid
result=""


RestartXServ() {

        sudo killall xinit
        sleep 5
        service hal restart
        . /usr/pluto/bin/Start_X.sh
        WriteTimeFile
}

WriteTimeFile() {

        if [ -f $timefile ]; then
                rm $timefile
        fi

        ctime=$(date +"%D %T")
        echo "$ctime" > $timefile
}

#First see if a timefile even exists
if [ ! -f $timefile ]; then
        #Clearly we have rebooted, so we need to restart the xserv to get the penmount working
        #Make sure that the LMCE services have been started and are currently running before we do this
        if [ -f $pidfile ]; then
                ps -A | grep $(cat $pidfile )>/dev/null
                err=$?
                if [ $err -eq 0 ]; then
                        RestartXServ
                fi
        fi
else
        #Otherwise, lets check to see if there has been a restart request since the last time we checked
        if [ -f $logfile ]; then

                #Search the logfile for the last restart request
                result=$(cat $logfile | grep "Got a reload command" | tail -n 1 | cut -f 2 )

                #Get epoch versions of the timestamp
                t1=$(date -d"$result" +%s)

                #Get epoch version of timestamp in timefile
                result=$(cat $timefile)
                t2=$(date -d"$result" +%s)

                if [ $t1 -gt $t2 ]; then
                        #We have a new request so restart the xserver
                        echo "Restarting X"
                        RestartXServ
                fi

        fi

fi

The key to fixing the problem on a router reload was restarting the hal daemon. Sadly the problem resurfaces on a reboot. The only way to fix it is to first let the system load, touch the screen (where the mouse pointer gets stuck to the left side of the window -- i.e. not working) then reload the xserv. After that, any subsequent reloads from the router works fine.

My next guess is then that the system doesn't know to listen on the serial port until it registers some sort of input first, then restarting hal (i presume) defaults to listening on that port and hence reloads work fine.

I can't figure out where or what in the hal daemon is changing when I first touch the screen that registers the serial port as active. If there were a way to force it to listen on the port (or inject an event to make it look like data came in on the port during init) that might work too...

Otherwise I might need to find a way to compile a kernel driver for the penmount to load at boot and I presume that will open the serial port by itself.

Thoughts?

EDIT: Forgot to attach files, they are there now

[JaseP]

There's almost got to be a kernel module for the driver, or a patch of the psmouse driver to do the same...

A common fix on systems is the setserial command, using it in the init script to prevent the psmouse driver from grabbing the ps2 serial port away from the touch driver. Most of the touch screen drivers are egalax derivatives. They mostly work the same way.  The newer the Ubuntu Linux version, the harder it is to get the touch screen drivers, connected to ps2 ports, to work,... up until a couple of fixes introduced from Ubuntu 10.10 on,... The udev rules for ps2 touchscreens weren't there, and the touch screen drivers all required a static serial port assignment. Udev would assign port IDs pretty much randomly, which screwed up the driver's ability to communicate. One time you'd boot with the device assigned to /dev/ttys04, and the next time it'd be /dev/ttys05,... And when you did finally get it recognized, you didn't have any calibration utility...

From your description of the problem,  it seems like the device is attaching only when the udev system realizes that there's something attached to that port assignment and grabs the right port the driver is meant to look at...

So there's probably one of two fixes for the issue,... Either use setserial in the init script to force the system to assign that serial port for touch driver use (maybe loading a kernel module to that process), or,... write a udev rule that recognizes that penmount device, and sets the right serial port environment for it...or some combination of the two...

By the way,... when you have a choice,... you always want a USB based controller chipset for touch screens in Linux. They're just easier (not always your choice, I know).