VinzTech

When running applications such as "unetbootin" or "tuxboot", you might get a blank window, instead of the actual application. Even when running it through "gksudo", or having a decent polkit running (e.g. inside Gnome). E.g.: This is because these application use MIT_SHM (Shared memory), which for some reason does not work when running graphical applications as root, while being logged in as a regular user in X. The solution is to disable MIT_SHM. E.g.: $ sudo QT_X11_NO_MITSHM=1 tuxboot

I have been struggling a few days now to get the fantastic "backports" project up and running on my machine (Arch Linux). Here's what I had to do: First, make sure you have backport checkout out, and pulled in all the required trees: # git clone git://git.kernel.org/pub/scm/linux/kernel/git/backports/backports.git # cd backports # ./devel/backports-update-manager  Here's what I found out: - Make sure your PYTHON_PATH does not contain funny stuff - devel/pycocci expects Python 2. Arch Linux defaults to Python 3. Use a virtualenv2, or change devel/pycocci's first line to "#!/usr/bin/env python2" - devel/pycocci crashes randomly when using different threads. I have hacked the script to make it single-threaded. EDIT: This seems to be because of Coccinelle 1.0.6. It does not occur using Cocinnele 1.0.4 - Coccinelle 1.0.6 does not work with all patches included in backports, you need version 1.0.4 with python support. (bin can be found here: ht...

This post is especially an outing of my joy now that I have finally resolved my issue debugging the STM32F4xx using OpenOcd. There exist many blogpost, tutorials, manuals about setting up an STM32F4Discovery + OpenOcd + gdb combo. I won't explain how to do it here, but it is an awesome, powerful and very cheap set-up. I chose the STM32F4Discovery board as the demo board for my course on embedded C, because it's cheap, full of features and has on-board debugging hardware. Writing a custom crt0, I ran into problems however. Stepping through the early init code, assembly instruction per assembly instruction, gdb would just lock-up, freeze, at random moments. Not always on the same instruction, not always at the same moment. I could also set a breakpoint on a certain line, then type 'c' to continue, and the debugger would never break, or even respond to a Ctrl+C. Furthermore, using st-util, from the texane github ( https://github.com/texane/stlink ) would work.It...

Another awesome project I've been busy with: I'm working for a company ( TASS ) that is specialized in embedded systems development and consultancy, and we have been working for quite some time on the PicoTCP project. This is a TCPI/IP stack, with a small footprint (so very suitable for embedded systems) and a very complete feature-set. You can check it out for yourself, since it's free: (GPL v2, but dual-licensed should you want to use for your company) https://github.com/tass-belgium/picotcp We have demos of the stack running under linux with a virtual ethernet device (VDE), and on a whole bunch of development boards, such as the STM32F4 boards, a Ti Stellaris Board, a PIC24, ... Check it out!

Ok, What have I been up to (technically speaking) ? Mostly, two things: The STM32F4-Discovery board which is a cheap and awesome developmentboard by STMicroelectronics boasting an ARM Cortex-M4 and an on-board JTAG-dongle so to speak: That opens a lot of interresting perspectives. But I'll talk about that in a next blogpost. The second thing is a combinations of the following technologies: GPS trackers + SMS + USB GSM Modem + Java + JavaScript + Google Maps. The idea is to have live tracking of about 20 GPS-trackers on a Google Map. This would serve as central intelligence center in a big real-life "hunting" game we're preparing at the moment to play with about 100 youngsters in the city of Tongeren. It will be called "Nacht van de Jacht", for those interrested :). The set-up I thought of was the following: We bought about 20 Chinese GPS trackers that are sending their position in an SMS to a central phone number every 3 minutes or so. These SMSes ar...

As I said before, I've been fiddling with finding the JTAG port on a Verbatim PowerBay NAS. I did not succeed (yet), but have had the chance to try out JTAGenum for Arduino. It is an excellent JTAG pin finder (and other things too). Check it out here: http://deadhacker.com/2010/02/03/jtag-enumeration/ The guy's description of the tool is: 1. Given a large set of pins on a device determine which are JTAG lines 2. Enumerate the Instruction Register to find undocumented functionality 3. be easy to build and apply It's really easy to use, but it did not manage to find the JTAG pins on my device. There might be another problem, such as: - JTAG is disabled once the device is powered up - The pins I'm fiddling with are not JTAG pins - I ruined the circuitry already - I'm just not trying hard enough Back to JTAGenum: Since Arduino IDE 1.0.1 , some things have changed and so the GIT version of JTAGenum would not compile as-is. Therefor the JTAGenum sketch had ...

Stuff I have been occupied with last few weeks, and that I'm planning to write my findings about: - Samsung CLP-300 Toner reviving: The Samsung CLP-300 is a nice color laser printer, except that it has programmed obsolesence. The toner and the printer contain an EEPROM memory, which accounts for the amount of pages printed and so. They become invalid after a predefined number of pages. (About 1000 i think). The good news is, that this is a simple I2C EEPROM memory, and that it can be re-programmed with a few tools. I have succesfully reprogrammed a few toners to behave like new ones last night, using an Arduino One and information lying around the Internet. Details will follow. - Verbatim PowerBay NAS hacking: My dad has a Verbatim PowerBay NAS lying around, which was pretty crippled in functionality and performance. That's why it's never really been used to do anything usefull. I tried to "break into" the system, software-wise, but couldn't find an e...