Ubuntu on Eee PC T101MT

[Plippo]

There has been a kernel update today, so there's a new version of the driver. It's linked in the first post, just install the new kernel driver for 2.6.32-22 after you have done the update (you don't have to uninstall the old driver, you also don't have to reinstall the calibrator, just download the new kernel driver and install it).

@mrspacklecrisp: I appreciate your excitement And don't worry, you haven't been as annoying as you think. I also noticed the bug in MyPaint and tried to work around it in the driver without luck. Can you give me a link to the corresponding bug report? If there is already a patch, maybe I can create a package with the new version.

@dusjagr: Good to know! I'm working on a small daemon that will allow things like two-finger right click, two finger scrolling, two finger zooming and two finger rotation in many applications without having to change them, but it's far from finished, so in the meantime, your workaround is quite nice.

[mrspacklecrisp]

Plippo, I think I love you. Really, your devotion astounds me.

Anyway, yes, the bug I'm referring to has been reported.https://gna.org/bugs/index.php?13622 It's the ghost line problem. Mypaint expects movement to come from a regular graphics tablet with which you move the cursor by hovering. It assumes the touchscreen is a mouse, and thus isn't happy.

Several users with special input devices (eg. capacitive touchscreens) have reported that MyPaint adds extra lines between the strokes.
This happens when MyPaint does not see any motion events between the strokes. It will interpolate position and pressure from the last motion event.

The workaround is in to either disable pressure information (in settings->pressure, set "mode for input devices" to "disabled") or to require some minimal pressure before painting (in settings->pressure, add a new control point to the "global pressure mapping" curve and drag it to (0.2, 0.0).

There were several attempts to fix this, the most recent one, reverted again in svn r330 since it caused problems on normal tablets. To find a good fix that works for everyone, someone has to analyze what GDK events you actually get on those devices and in what order.

On another note, I have Karmic on a usb stick, so I can take a look at the fn keys issue. Plippo, you may want to direct me to make sure I check in the right places. I can also save some helpful things (say, to replace the events in lucid) to an sd card.

[Confuseling]

Hi Plippo;

Thanks for your work on this. My t101mt arrived today, so I'm going to be dual booting it, though I tend to use Debian these days, and would prefer to run that if I can.

Any chance you could make the sources available so we can have a crack at building these ourselves?

[Plippo]

Of course you can get the sources, it's all free software, so it's your right to get them.

The kernel driver is provided by ENAC, you find sources and building instructions there:
- http://www.lii-enac.fr/en/projects/s...nux-howto.html contains general build instructions
- http://www.lii-enac.fr/en/projects/s...h-devices.html contains the specific driver - look for "eGalax"
For the package I've modified the driver a bit, one patch has already been sent to ENAC but not yet been accepted, the other one has not been sent there yet, I'll do that later and also post the patches then, I don't have them available right now.

The calibrator is basically this one: http://www.freedesktop.org/wiki/Soft...put_calibrator - I've just modified it a bit so it directly writes the calibration to the xorg.conf.d path, but depending on which version of Debian you use, I don't now if they already use this mechanism or still one of the older udev or hal or Xorg.conf mechanisms. In any case, I'll also post my modifications for this later.

[Confuseling]

...In any case, I'll also post my modifications for this later.

Don't let me rush you - all this is very much appreciated. I may well end up installing Ubuntu on it anyway until the useful stuff flows up to Debian or the kernel - I'll have a play, and decide what looks more feasible.

Cheers.

[riccardo.depalo]

Well, I must thank Plippo for his work: his driver is the only one which works without any problem. I tried also the driver from eeti (I found it here: http://home.eeti.com.tw/web20/eGalax...inuxDriver.htm ) but it doesn't support drawing with Gimp or Xournal, for instance. Anyway, the driver (I tried the beta one) comes with a calibration tools which supports 4, 9 and 25 points of calibration, and also a drawing test. Maybe it could be used to improve pen response of the t101mt.

[Plippo]

If I don't do it now, I have to do it later, so why not do it now?

So, attached to this post is an archive with the patches for the kernel module and, for your convenience, also the already patched module file. So you can either download the original hid-egalax.c from the ENAC homepage and apply the two patches or simply use the hid-egalax.c from the archive.

The source code of the modified calibrator can be downloaded from http://www.philmerk.de/dwl/deb/egala...0.2-src.tar.gz

[Plippo]

Hi riccardo,

yes it would really be nice if we had 9 or 25 point calibration, but unfortunately the EETI driver is closed source, so we can't improve it or use it to improve our multitouch driver. So this shows again why closed source software is a bad, bad thing.

[Plippo]

@mrspacklecrisp: Thanks for the love, you can never have enough of it.

And thanks for your offer to do some experiments in Karmic. The thing that would be of greatest interest is whether the eeepc-laptop module or any other acpi module is used in Karmic. To test this, you can go to a terminal and call

lsmod

and post the output here.

Another thing you could post is the output of the command

cat /proc/acpi/video/VGA/LCDD/brightness

About MyPaint: I'll try to compile that SVN Revision 330 or extract the patch from there, but then it will probably only work for resistive screens and no longer for Wacom tablets, so if you also want to use a Wacom tablet with the EeePC, that won't work. But I'll try it when I find the time.

[mrspacklecrisp]

Suspicions about lucid being darker are right. In Karmic the screen is lit up like the sun. If I'm told what commands/files can help, it'll only take a couple of minutes.

ubuntu@ubuntu:~$ lsmod
Module                  Size  Used by
binfmt_misc             8356  1 
ppdev                   6688  0 
lp                      8964  0 
parport                35340  2 ppdev,lp
snd_hda_codec_realtek   203328  1 
snd_hda_intel          26920  2 
snd_hda_codec          75708  2 snd_hda_codec_realtek,snd_hda_intel
snd_hwdep               7200  1 snd_hda_codec
snd_pcm_oss            37920  0 
snd_mixer_oss          16028  1 snd_pcm_oss
snd_pcm                75296  3 snd_hda_intel,snd_hda_codec,snd_pcm_oss
snd_seq_dummy           2656  0 
snd_seq_oss            28576  0 
snd_seq_midi            6432  0 
arc4                    1660  2 
ecb                     2524  2 
snd_rawmidi            22208  1 snd_seq_midi
iptable_filter          3100  0 
snd_seq_midi_event      6940  2 snd_seq_oss,snd_seq_midi
ip_tables              11692  1 iptable_filter
dm_crypt               12928  0 
x_tables               16544  1 ip_tables
snd_seq                50224  6 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
snd_timer              22276  2 snd_pcm,snd_seq
snd_seq_device          6920  5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
uvcvideo               59080  0 
ath9k                 258744  0 
mac80211              181236  1 ath9k
led_class               4096  1 ath9k
videodev               36736  1 uvcvideo
v4l1_compat            14496  2 uvcvideo,videodev
snd                    59204  16 snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_seq_oss,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
joydev                 10272  0 
ath                     8060  1 ath9k
psmouse                56180  0 
cfg80211               93052  3 ath9k,mac80211,ath
atl1c                  30880  0 
serio_raw               5280  0 
soundcore               7264  1 snd
snd_page_alloc          9156  2 snd_hda_intel,snd_pcm
eeepc_laptop           13936  0 
squashfs               22912  1 
aufs                  149420  1 
nls_iso8859_1           3740  2 
nls_cp437               5372  2 
vfat                   10716  2 
fat                    51452  1 vfat
dm_raid45              84228  0 
xor                    15620  1 dm_raid45
fbcon                  36640  72 
tileblit                2460  1 fbcon
font                    8124  1 fbcon
bitblit                 5372  1 fbcon
softcursor              1756  1 bitblit
usbhid                 38208  0 
i915                  221064  3 
drm                   159584  3 i915
i2c_algo_bit            5760  1 i915
usb_storage            52544  2 
intel_agp              27484  2 i915
agpgart                34988  2 drm,intel_agp
video                  19380  1 i915
output                  2780  1 video
ubuntu@ubuntu:~$ cat /proc/acpi/video/VGA/LCDD/brightness
levels:  0 3 5 7 9 11 13 16 19 23 26 30 33 37 41 46
current: 16

Note: Brightness was at the highest level I could put it at with the fn keys. It was much brighter than the maximum in Lucid.