stan.distortion- Is this the xorg.conf input section you are using?
I can see output from the touchpad on (/dev/input/)event2, mice, and mouse0, but nothing from the touchscreen.

Is it necessary to reboot after mucking about with lifebook.c, or does
suffice to reinitialize whatever needs to be reinitialized, (after rebuilding the modules, of course)?

Azizcoos,

Your box, like CF-18, has an active multiplexing controller, but, unlike CF-18 we only detect one PS/2 device connected to it. It coudl be that your touchpad is not PS/2 but a serial one. Have you tried using the driver from the link below?

http://stz-softwaretechnik.com/~ke/t.../p-series.html

Has anyone had any success with either the cf-18 or cf-29? If so, could you post some info.
Thanks,
Stan

Ok, for CF-29: as I said it could be that it has a serial touchpad instead of PS/2 one. In that case please try grabbing fujitsu-serial-ts.patch and a precompiled inputattach binary from

http://www.kernel.org/pub/linux/kernel/people/dtor/

if inputattach does not work for you you can compile it from inputattach.c and serio-ids.h that you can find at the link above as well.

Once you patched your kernel you need to load 8250 (or 8250_pnp) module, serport module and fujitsu_ts module. then do

inputattach -fjt /dev/ttyS0

It could be that the touchscreen is not at ttyS0 but ttyS1 (or 2, 3, 4). You may have to play with setserial to setup the ports.

As I said I am working off the development tree so the patch shoudl apply to 2.6.20 and 2.6.21-rc* but it may give you some fuzz and fits. Sorry about that.

Thanks to Dimitri and Stan, I got now my CF-18 touchscreen and touchpad working. Touch screen is not calibrated but what is most important, touchpad is working. I did applied patch from Dimitri and it worked out, he can propably give some details how it was made.

Next thing is touchscreen calibration. How do I calibrate the screen?

Hi all. I have had a cf-18 for quite a while now and have been wracking my brain trying to get the touchscreen and pad working right. I upgraded to suse 10.2, and am trying to straighten out the mess the default kernel made with my touchpad/screen. I'm glad that PPaFin succeeded with his. I am a bit disheartened to hear that it requires patching the module. I was hoping it just required a good xorg configuration.

I don't know if this is still relevant, but my system does differentiate the signals from the touchscreen and the touchpad. The touchscreen input goes to /dev/input/mouse0 while the touchpad input goes to /dev/input/mouse1. I'd hoped I could create two different input sections in xorg.conf that dealt with the touchscreen signals as "absolute" and the touchpad signals as "relative" but no luck. As a matter of fact, nothing I do to the config seems to make any difference at all.

PPaFin, if you could detail exactly what you've done (what modules you've patched, what modules you've loaded, the xorg.conf you use, etc.) that would be really appreciated.

As for calibration, I believe I downloaded such a program that was bundled with conan's X touchscreen drivers at conan.de for the lifebook B series. Instructions on using it are included I think.

Well, come to think of it you can avoid patching lifebook.c if you have CF-18. You just need to issue the following command after loading psmouse module:

echo -n 'bare' > /sys/bus/serio/devices/serioX/protocol

where serioX is serio port your touchpad is connected to.

Well, the CF-18 doesn't use serial ports for the touchscreen. Both the touchscreen and touchpad (as far as I know) share the psaux device.

The serio part is a reference to the type of communications, a bit like char or block devices in the /dev part of the file system.
Try the command Dtor posted. The trouble your having isn't due to interpretations of the information from the mouse device nodes (which is why making changes in xorg.conf has no effect), it is due to the physical setting of the hardware controller. In effect a switch has been flipped in the controller to tell it to give absolute coordinates.
The command from Dtor above will (in effect) flip the switch back to give relative output.
Hope that helps. If you have the time, it would be worth trying the patch. The touchscreen is a 'must have' on a laptop after using one for a while. Its not that hard to apply it and compile the kernel, almost everything is automated and there are plenty of good howto's on the net for every distro.
Cheers,
Stan

OK, I feel pretty confident about changes for CF-18 and CF-28 so they are scheduled for 2.6.22. I am very interested in CF-29 and whether fujitsu-serial-ts.patch works with it.

Thanks!

Ok I will try that. For now, though, I have the touchscreen and touchpad working. (I had a small problem with the xorg config file labeling both of them as corepointers - I just didn't catch it) I am using the evdev driver for the touchpad, with Option "Mode" "Relative" enabled in the xorg.config. It works now with its relative axises, but is very sensitive to movement, and doesn't give me the Emulate3Buttons feature that I prefer.

I want to still try the command dtor offered with another mouse driver, in hopes of getting the sensitivity down. Stan, perhaps this was mentioned earlier in the topic, but what is the reason for the module patch?

Last thing. To get the touchscreen calibrated properly (well, a semblance of it) I used the calibrator from the conan.de package. There is an error in it, so to fix it requires compiling the source package and editing the file ev_calibrate.c at line 290 to read
CURSORDIRECTORY"./empty_cursor.xbm",

After compiling it and running calibrate.sh from the same directory, the program works.

Edited April 24th, 2007

Hey Guys,

Just a quick Reader's Digest story here. On March 29th, my friend that is a missionary to Haiti stopped by my shop. I have been helping him look at Toughbooks for about two months since his R32 ThinkPad I sold him was stolen on his last trip there(it was stolen here in the states when the plane landed). We were talking and I showed him some CF-28s that I had found that had factory built-in GPS units. I was whining that it was the one that I had been looking for forever and I had spent all my "extra" cash on the unit that was being delivered that afternoon. He looked at me kinda funny and asked, "Do you think that you could have it ready to leave for Haiti in 36 hours?" I replied, "Of course!" He then took out his checkbook and wrote out a check and gave it to me. He then said, "In that case, order two of them, one for you, and one for me. Call when your ready for me to pick this one up, and I'll see you when I get back."

Needless to say I have been busy getting GPSs working since I have both of them to play with until he gets back. The funny thing is, it was easier to figure out how to get them working under Linux that Windows XP. Short story is that XP has a bug in it that sees the constant stream of data coming from the serial port as a "Microsoft Serial BallMouse". Remember when I said that if you get the newer version of lbtouch your cursor would jump around like it was on crack? That was NOTHING, NOTHING, compared to what XP did with it. It saw the GPS data stream as mouse coorodinates changing....and tried to keep up. Imagine that.

I am going to take off 9 days starting this Saturday, so I will have some time to mess around with applying these kernel patches and recompiling. If I can manage to get a working 2.6 touchscreen to go along with my working GPS, I'll be the happiest guy in town. At least for a day or so... :-)

James

How have you done? Can you link plz?

I am sorry guys, I have been quite busy lately (I will try to arrange DMI information of those new Toughbooks for Stan and Dimitri).

What I did for this CF-18 was following;

1. Installed kubuntu 6.10
2. Everything was working nicely, until kernel upgrade (2 month ago) which messed up my touchpad (it started to act as relative pointer).
3. I reached Stan for help here and gave also some emails to Dimitri.
4. Guys proposed multiple solutions and I tried couple of those.
5. I got kernel 2.6.20 sources, reminded myself how to compile kernel and modules. It was good exercise anyhow, but in my normal business days I have no such time in my hands often.
6. What worked (as quick workaround) was patch to lifebook.c and re-compile kernel modules (to produce new psmouse.ko file).
7. Lifebook.c patch disabled only detection strings for CF-18 model and let system load non touch screen version of psmouse.ko (?)
8. More professional solution was to patch psmouse-base.c as proposed by Dimitri.

If some-one is still interested actual patch files, I can find them here or maybe Dimitri can give us some directions how to make this in right way. I personally think that it would be easier to disable just lifebook.c originated detection string of CF-18 model, but this actual psmouse-base.c patch did work and leaves touchpad and touchscreen working.

As we have also multiple toughbooks in our office and I try to evaluate their suitability in various linux based projects, this forum has been great help for me. Currently I have CF-18, CF-29 running linux and CF-Y5, CF-19 scheduled to be evaluated for suitability.

Current challenge is in build GPRS/EDGE/3G/HSDPA modules of those laptops, generally CF-18 GPRS and CF-29 EDGE modules are working nicely. CF-Y5 has Novatel wireless 3G/HSDPA module which is still under investigation (it is no PCCard, it's in build and interfaced by internal USB).

I have faced this same situation years back with CF-28 and Windows 2000, m$ had knowledge base article for that in those days. It was all about bug in mouse driver, which used incorrectly nmea strings for positioning of mouse pointer

I am happy to play with Linux on these Toughbooks, rather than windows (I think game consoles and IT-systems should be from different providers