Nokia N810 GPS fix solution

The Nokia N810 with its internal GPS is a device that should fit well for navigation purposes. Below you can see it running navit. In my opinion this is the best available navigation software for the N810. Navit is the main reason for me for wanting the GPS unit in the N810 to be reliable. Get the svn version from here: Gerrit’s Techie Stuff. Some hints on getting navit to run on the N810 can be found on

N810 running Navit
N810 running Navit

Sadly the internal GPS unit of the N810 is known to be very problematic and occasionally hardly gets a fix. The suspect seems to be the gpsdriver. It writes i.e. the almanac data to /var/lib/gps/nvd_data. This file seems to be corrupt sometimes when the gpsdriver writes it. Probably the problem occurs when the GPS unit has no fix. The bug entry on this topic on inspired me to write a small wrapper script for the gpsd that copies a backup to /var/lib/gps/nvd_data and watches the gps-daemon. When the GPS unit has a fix it creates a new backup copy of the nvd_data file for later use. Since I use the script, acquiring a GPS fix never took me more than five minutes anymore (without AGPS, with AGPS it is faster, of course).

If you also want to give my script a try, install bash via program manager and download my script from here: gpsd wrapper script for n810 (1402)

The most convenient way to get the script to the n810 and to perform the necessary changes is to log into the tablet using ssh. How to get ssh to run on your internet tablet is described here: internettabletblog

Log into your tablet using ssh and rename the original gpsd.

-sh-3.2# mv /usr/sbin/gpsd /usr/sbin/gpsd.orig

Then copy my script to the tablet (replace XXX.XXX.XXX.XXX with the ip of your tablet):

scp gpsd root@XXX.XXX.XXX.XXX:/usr/sbin/

Afterwards fix the permissions of the script on the tablet:

-sh-3.2# chmod a+rx /usr/sbin/gpsd

That’s all, now you should have a much better working GPS unit in your N810. Since I only have one device to test this on, please leave a comment if this also worked for you.



1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

nvidia-drivers-180.60 is working !!!

After I have tested the nvidia-drivers-185.18.08 beta I searched the web for a new one and I found it yesterday on This version seems to run well on my Dell M4300. For all using Gentoo I  made an ebuild:ebuild:nvidia-drivers (993)

# mkdir /usr/local/portage/
# cd /usr/local/portage/
# tar -xpvf nv-portage.tar

Be sure that PORTDIR_OVERLAY=”/usr/local/portage” is in your /etc/make.conf, then you can emerge it:
# ACCEPT_KEYWORDS=”~amd64″ emerge =nvidia-drivers-180.60 -1

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

nvidia-drivers-185.18.08 beta is not working

Because of the problem with a good, fast and working nvidia driver, I’ve tested the actual beta driver from Nvidia. After installing it on my system (Dell M4300, with zen-sources 2.6.29-r3 running fine on it) everything seems to be fine. But there was the first fault. The character set in the Terminal swaps. The next serious error was restarting the X-Server or crashing the whole system while surfing with Epiphany. So I downgraded my system to nvidia-drivers-180.29 again as mentioned here. It seems Nvidia has to do some work on it, still we have another working release. Stay tuned, we will test it and tell you our experience with any new driver we found from Nvidia.

LG Thorsten

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

upgrade experiences with nvidia-drivers-180.51

Today, during my regular upgrades, I upgraded my gentoo boxes to nvidia-drivers-180.51, which is the latest version in portage. Since the error in dmesg

bluetooth-apple[10079] general protection ip:7fec32b2cebe sp:7fff3dc7ff00 error:
0 in[7fec32ad6000+a5000]

I mentioned in CTRL-C and CTRL-Z not working after nvidia-driver upgrade and high CPU usage and saw in later releases still occurs and therefore I expect more problems, I downgraded to nvidia-drivers-180.29 again.  Since I upgraded  to xorg-server-1.5.3 in the meanwhile, I am sure now, that the problems are not related to incompatibilities with the older xorg release I ran before.

Did anyone monitor the driver release any closer? How were your experiences?


1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

zen-sources-2.6.29-r3 with tuxonice

Some days ago Thorsten wrote TuxOnIce is broken in 2.6.29-zen1. Well, I wanted to have a 2.6.29 kernel, because the iwl3945 driver in 2.6.29 works the first time with a useable speed for me.  So I no longer need the hack for the ipw3945 driver I published on ipw3945 unter linux-kernel 2.6.27.

Today I managed to patch the brand new zen-sources-2.6.29-r3 with tuxonice. This was necessary since I need tuxonice on my notebook and tuxonice has been removed permanently from zen-sources. Suspend to RAM works at least on my Dell Precison M65 as well as suspend to disk does.

Since 2.6.29-zen3 has not been released as patch yet, you may download the patch I created from git here: Zen-Sources 2.6.29-r3 patch (885) Then apply the patch to the vanilla kernel sources available from Afterwards download the tuxonice patch from and apply it to the kernel sources with the zen patchset already applied.

For Gentoo users there is a more easy way:

Download my modified overlay  (including the patches) from zen-sources-toi-2.6.29-r3.tar.bz2 (967) and extract it in /usr/local/portage. Be sure to include the following line in your /etc/make.conf:


Then emerge zen-sources-toi and buld the kernel as you wish.

The files above are not from an official zen release. Tuxonice is not officially supported in zen-sources. So If you’re using the files above, don’t report any bugs to

For my Precision M65 I used the following kernel config: (936)

If you want to use all channels with the iwl3945 driver change the line

static char *ieee80211_regdom = “US”;


static char *ieee80211_regdom = “JP”;

in the file net/wireless/reg.c.

For more information on the zen-sources patchset see

best regards


1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)

SIV-1.0 released

I’m proud to present a new version of SIV, the stereoscopic jps viewer for Linux I published first in SIV – a stereoscopic jps viewer for Linux. SIV is  capable of displaying jps stereo images in different stereo modes. I tried it in fullscreen/windowed mode with anaglyphic and quad buffered stereo mode. Different output devices I tested were my Vuzix VR920 and relevator shutter glasses  on Nvidia Quadro FX 350M and FX 3400. It should also work with an Ati FireGL, but I never tried this. The new version fixes some segfaults, that occured when needed libraries were missing. Furthermore I added a slideshow mode and the possibility to specify a default distance and therefore the possibility to set a default zoom factor. The coolest feature I added to the viewer is the headtracking support for the Vuzix VR920 HMD. For this you will need (in addition to the device 😉 ) the headtracking driver, I published in VR920 headtracking driver for Linux. You may watch  a video (demo_siv_headtracking_vr920) of me using the viewer with enabled headtracking to view some photos of NASA’s Apollo missions  downloadable from nvidia forums.

Usage changes:

New commandline options are:

  • -s                                         Enable slideshow mode
  • -t                                         delay Delay for slideshow mode,
    defaults to 10 sec
  • -d                                         distance (values between 0.0 and
    1.0 are senseful)

These new options control the new headtracking support:

  • –vr920                             enable headtracking support for the vuzix vr920 hmd
  • -m                                        non default multicast ip
  • -p                                         non default multicast port

For detailed information on the -m and -p options see the article about the headtracking driver : VR920 headtracking driver for Linux. They have the corresponding meaning. If you want to try out the headtracking, enable the headtracking support with the –vr920 option. You may want to specify the distance to the image with the -d option. Values between 0.25 and 0.5 seem to have a nice effect when using the headtracking. When headtracking is active you may press Z to set the zero view so that with your current head position you will see the center of the image. If you want to manipulate the view manually press 2 to enable the trackball manipulator and use i.e. the right moue button and up or down movements (with the mouse) to control the distance (zoom in or out). Afterwards press 1 to activate headtracking again.

For more usage documentation see: SIV – a stereoscopic jps viewer for Linux


I decided to publish the viewer under the creative common noncommercial license. Make sure you have the OpenSceneGraph library installed before trying to compile or run the viewer. You may download the full source (Eclipse Project) from here: SIV-1.0 source (1242) or an x86_64 binary from here:  SIV-1.0 x86_64 binary (1256). More binary formats may be available in the future. Fore the x86_64 binary you’ll need OpenSceneGraph-2.8.

Binaries for most linux distributions can be downloaded from  Instructions on how to build OpenSceneGraph on gentoo can be found on PlopByte.

Building from source:

For building from source unpack the zip and cd to siv/Release and run make.


If you like the viewer, feel free to link to To request commercial licenses contact us at info(at) Well, if you just want to support our work on use the donate button 😉

best regards and enjoy the 3rd dimension


1 Star2 Stars3 Stars4 Stars5 Stars (1 votes, average: 5.00 out of 5)

Upgrade to nvidia-drivers-180.41

Today I upgraded my Gentoo systems to nvidia-drivers-180.41.  I had bad news to report the last two times I upgraded nvidia-drivers in nvidia-drivers-180.37 causing strange segfaults and CTRL-C and CTRL-Z not working after nvidia-driver upgrade and high CPU usage.

Well, since I had bad news before I also want to tell you about the good news. I did not experience any problems yet with nvidia-drivers-180.41. I run Gentoo with kernel 2.6.28-zen4 #1 SMP PREEMPT on x86_64 and tried with Quadro FX3400 and Quadro FX350M. Even tuxonice works on my  Dell Precision M65 with the driver. In the past I sometimes had problems with suspending and some nvidia-driver versions.

Well, not everyone is as lucky as I am. Some cards seem not to be recognized anymore by the driver. You can read about this problem on 180.41 — no more support for 6600?

I’m quite happy to have a new working driver after the last two upgrade attempts.

Edit: Sadly I was wrong in this post. nvidia-drivers-180.41 also causes segfaults as well as 180.44 does. These segfaults occur during heavy CPU/MEM usage. (In my case running 6 instances of mencoder simultaneously on 4 cores). After downgrading again to 180.29 everything returned to normality.


1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)