"In het verleden behaalde resultaten bieden geen garanties voor de toekomst"
About this blog

These are the ramblings of Matthijs Kooijman, concerning the software he hacks on, hobbies he has and occasionally his personal life.

Most content on this site is licensed under the WTFPL, version 2 (details).

Questions? Praise? Blame? Feel free to contact me.

My old blog (pre-2006) is also still available.

Sun Mon Tue Wed Thu Fri Sat
18 19 20 21 22 23 24
25 26 27 28      
Tag Cloud
Powered by Blosxom &Perl onion
(With plugins: config, extensionless, hide, Markdown, macros, breadcrumbs, calendar, directorybrowse, entries_index, feedback, flavourdir, include, interpolate_fancy, listplugins, menu, pagetype, preview, seemore, storynum, storytitle, writeback_recent, moreentries)
Valid XHTML 1.0 Strict & CSS
Calculating a constant path basename at compiletime in C++

Tags: Nerd, Softare, Optimization, C++, Templates

In some Arduino / C++ project, I was using a custom assert() macro, that, if the assertion would fail show an error message, along with the current filename and line number. The filename was automatically retrieved using the __FILE__ macro. However, this macro returns a full path, while we only had little room to show it, so we wanted to show the filename only.

Until now, we've been storing the full filename, and when an assert was triggered we would use the strrchr function to chop off all but the last part of the filename (commonly called the "basename") and display only that. This works just fine, but it is a waste of flash memory, storing all these (mostly identical) paths. Additionally, when an assertion fails, you want to get a message out ASAP, since who knows what state your program is in.

Neither of these is really a showstopper for this particular project, but I suspected there would be some way to use C++ constexpr functions and templates to force the compiler to handle this at compiletime, and only store the basename instead of the full path. This week, I took up the challenge and made something that works, though it is not completely pretty yet.

Working out where the path ends and the basename starts is fairly easy using something like strrchr. Of course, that's a runtime version, but it is easy to do a constexpr version by implementing it recursively, which allows the compiler to evaluate these functions at compiletime.

For example, here are constexpr versions of strrchrnul(), basename() and strlen():

 * Return the last occurence of c in the given string, or a pointer to
 * the trailing '\0' if the character does not occur. This should behave
 * just like the regular strrchrnul function.
constexpr const char *static_strrchrnul(const char *s, char c) {
  /* C++14 version
    if (*s == '\0')
      return s;
    const char *rest = static_strrchr(s + 1, c);
    if (*rest == '\0' && *s == c)
      return s;
    return rest;

  // Note that we cannot implement this while returning nullptr when the
  // char is not found, since looking at (possibly offsetted) pointer
  // values is not allowed in constexpr (not even to check for
  // null/non-null).
  return *s == '\0'
      ? s
      : (*static_strrchrnul(s + 1, c) == '\0' && *s == c)
        ? s
        : static_strrchrnul(s + 1, c);

 * Return one past the last separator in the given path, or the start of
 * the path if it contains no separator.
 * Unlike the regular basename, this does not handle trailing separators
 * specially (so it returns an empty string if the path ends in a
 * separator).
constexpr const char *static_basename(const char *path) {
  return (*static_strrchrnul(path, '/') != '\0'
      ? static_strrchrnul(path, '/') + 1
      : path

/** Return the length of the given string */
constexpr size_t static_strlen(const char *str) {
  return *str == '\0' ? 0 : static_strlen(str + 1) + 1;

So, to get the basename of the current filename, you can now write:

constexpr const char *b = static_basename(__FILE__);

However, that just gives us a pointer halfway into the full string literal. In practice, this means the full string literal will be included in the link, even though only a part of it is referenced, which voids the space savings we're hoping for (confirmed on avr-gcc 4.9.2, but I do not expect newer compiler version to be smarter about this, since the linker is involved).

To solve that, we need to create a new char array variable that contains just the part of the string that we really need. As happens more often when I look into complex C++ problems, I came across a post by Andrzej Krzemieński, which shows a technique to concatenate two constexpr strings at compiletime (his blog has a lot of great posts on similar advanced C++ topics, a recommended read!). For this, he has a similar problem: He needs to define a new variable that contains the concatenation of two constexpr strings.

For this, he uses some smart tricks using parameter packs (variadic template arguments), which allows to declare an array and set its initial value using pointer references (e.g. char foo[] = {ptr[0], ptr[1], ...}). One caveat is that the length of the resulting string is part of its type, so must be specified using a template argument. In the concatenation case, this can be easily derived from the types of the strings to concat, so that gives nice and clean code.

In my case, the length of the resulting string depends on the contents of the string itself, which is more tricky. There is no way (that I'm aware of, suggestions are welcome!) to deduce a template variable based on the value of an non-template argument automatically. What you can do, is use constexpr functions to calculate the length of the resulting string, and explicitly pass that length as a template argument. Since you also need to pass the contents of the new string as a normal argument (since template parameters cannot be arbitrary pointer-to-strings, only addresses of variables with external linkage), this introduces a bit of duplication.

Applied to this example, this would look like this:

constexpr char *basename_ptr = static_basename(__FILE__);
constexpr auto basename = array_string<static_strlen(basename_ptr)>(basename_ptr); \

This uses the static_string library published along with the above blogpost. For this example to work, you will need some changes to the static_string class (to make it accept regular char* as well), see this pull request for the version I used.

The resulting basename variable is an array_string object, which contains just a char array containing the resulting string. You can use array indexing on it directly to access variables, implicitly convert to const char* or explicitly convert using basename.c_str().

So, this solves my requirement pretty neatly (saving a lot of flash space!). It would be even nicer if I did not need to repeat the basename_ptr above, or could move the duplication into a helper class or function, but that does not seem to be possible.

0 comments -:- permalink -:- 21:33
Automatically remotely attaching tmux and forwarding things

Tags: Software, Linux, Tmux, Forwarding, Remote, Mutt, Script, SSH

GnuPG logo

I recently upgraded my systems to Debian Stretch, which caused GnuPG to stop working within Mutt. I'm not exactly sure what was wrong, but I discovered that GnuPG version 2 changed quite some things and relies more heavily on the gpg-agent, and I discovered that recent SSH version can forward unix domain socket instead of just TCP sockets, which allows forwarding a gpg-agent connection over SSH.

Until now, I had my GPG private keys stored on my server, Tika, where my Mutt mail client also runs. However, storing private keys, even with a passphrase, on permanentely connected multi-user system never felt quite right. So this seemed like a good opportunity to set up proper forwarding for my gpg agent, and keep my private keys confined to my laptop.

I already had some small scripts in place to easily connect to my server through SSH, attach to the remote tmux session (or start it), set up some port forwards (in particular a reverse port forward for SSH so my mail client and IRC client could open links in my browser), and quickly reconnect when the connection fails. However, once annoyance was that when the connection fails, the server might not immediately notice, so reconnecting usually left me with failed port forwards (since the remote listening port was still taken by the old session). This seemed like a good occasion to fix that as wel.

The end result is a reasonably complex script, that is probably worth sharing here. The script can be found in my scripts git repository. On the server, it calls an attach script, but that's not much more than attaching to tmux, or starting a new session with some windows if no session is running yet.

The script is reasonably well-commented, including an introduction on what it can do, so I will not repeat that here.

For the GPG forwarding, I based upon this blogpost. There, they suggest configuring an extra-socket in gpg-agent.conf, but I've found that gpg-agent already created an extra socket (whose path I could query with gpgconf --list-dirs), so I didn't use that extra-socket configuration line. They also talk about setting StreamLocalBindUnlink to clean up a lingering socket when creating a new one, but that is already handled by my script instead.

Furthermore, to prevent a gpg-agent from being autostarted by gnupg serverside (in case the forwarding fails, or when I would connect without this script, etc.), I added no-autostart to ~/.gnupg/gpg.conf. I'm not running systemd user session on my server, but if you are you might need to disable or mask some ssh-agent sockets and/or services to prevent systemd from creating sockets for ssh-agent and starting it on-demand.

My next step is to let gpg-agent also be my ssh-agent (or perhaps just use plain ssh-agent) to enforce confirming each SSH authentication request. I'm currently using gnome-keyring / seahorse as my SSH agent, but that just silently approves everything, which doesn't really feel secure.

0 comments -:- permalink -:- 16:46
Running Ruby on Rails using Systemd socket activation

Ruby on Rails logo

On a small embedded system, I wanted to run a simple Rails application and have it automatically start up at system boot. The system is running systemd, so a systemd service file seemed appropriate to start the rails service.

Normally, when you run the ruby-on-rails standalone server, it binds on port 3000. Binding on port 80 normally requires root (or a special capability enabled for all of ruby), but I don't want to run the rails server as root. AFAIU, normal deployments using something like Nginx to open port 80 and let it forward requests to the rails server, but I wanted a minimal setup, with just the rails server.

An elegant way to binding port 80 without running as root is to use systemd's socket activation feature. Using socket activation, systemd (running as root) opens up a network port before starting the daemon. It then starts the daemon, which inherits the open network socket file descriptor, with some environment variables to indicate this. Apart from allowing privileged ports without root, this has other advantages such as on-demand starting, easier parallel startup and seamless restarts and upgrades (none of which is really important for my usecase, but it is still nice :-p).

Making it work

To make this work, the daemon (rails server in this case) needs some simple changes to use the open socket instead of creating a new one. I could not find any documentation or other evidence that Rails supported this, so I dug around a bit. I found that Rails uses Rack, which again uses Thin, Puma or WEBrick to actually set up the HTTP server. A quick survey of the code suggests that Thin and WEBrick have no systemd socket support, but Puma does.

I did find a note saying that the rack module of Puma does not support socket activation, only the standalone version. A bit more digging in my Puma version supported this, but it seems that some refactoring in the 3.0.0 release (commit) should allow Rack/Rails to also use this feature. Some later commits add more fixes, so it's probably best to just use the latest version. I tested this succesfully using Puma 3.9.1.

One additional caveat I found is that you should be calling the bin/rails command inside your rails app directory, not the one installed into /usr/local/bin/ or wherever. It seems that the latter calls the former, but somewhere in that process closes all open file descriptors, losing the network connection (which then gets replaces by some other file descriptor, leading to the "for_fd: not a socket file descriptor" error message).

Setting this up

After setting up your rails environment normally, make sure you have the puma gem installed and add the following systemd config files, based on the puma examples. First, /etc/systemd/system/myrailsapp.socket to let systemd open the socket:

Description=Rails HTTP Server Accept Sockets


# Socket options matching Puma defaults



Then, /etc/systemd/system/myrailsapp.service to start the service:

ExecStart=/home/myuser/myrailsapp/bin/rails server puma --port 80 --environment production


Note that both files should share the same name to let systemd pass the socket to the service automatically. Also note that the port is configured twice, due to a limitation in Puma. This is just a minimal service file to get the socket activation going, there are probably more options that might be useful. This blogpost names a few.

After creating these files, enable and start them and everything should be running after that:

$ sudo systemctl enable myrailsapp.socket myrailsapp.service
$ sudo systemctl start myrailsapp.socket myrailsapp.service
1 comment -:- permalink -:- 10:35
Raspberry pi powerdown and powerup button

Tags: RaspberryPi, Hardware, Software, Linux, Electronics

Raspberry Pi Zero

TL;DR: This post describes an easy way to add a power button to a raspberryp pi that:

  • Only needs a button and wires, no other components.
  • Allows graceful shutdown and powerup.
  • Only needs modification of config files and does not need a dedicated daemon to read GPIO pins.

There are two caveats:

  • This shuts down in the same way as shutdown -h now or halt does. It does not completely cut the power (like some hardware add-ons do).
  • To allow powerup, the I²C SCL pin (aka GPIO3) must be used, conflicting with externally added I²C devices.

If you want to get started right away, skip to the "Setting this up" section below.

Powering down your pi

To gracefully power down a Raspberry Pi (or any Linux system), it has to be shut down before cutting power. On the rpi, which has no power button, this means using SSH to log in and running a shutdown command. Or, as everybody usually does, cutting the power and accepting the risk of data corruption.

Googling around shows a ton of solutions to this problem, where a button is connected to a GPIO pin. The GPIO pin is monitored and when it changes, a shutdown command is given. However, all of these seem to involve a custom daemon, usually written in Python, to monitor the GPIO pin. A separate daemon just for handling the powerbutton does not sound great, and if it is not apt-get installable, it seems too fragile for my tastes. Ideally, this should be handled by standard system components only.

Another thing is that a lot of these tutorials recommend wiring a pullup or pulldown resistor along with the button, while the raspberry pi has builtin pullups and pulldowns on all of its I/O pins which can just as easily be used.

Shutdown using systemd-logind

Having a shutdown button connected to a GPIO pin is probably not so uncommon, so I Googled some more. I came across the 70-power-switch.rules udev file from systemd-logind, which adds the "power-switch" tag to some kernel event sources representing dedicated power buttons. Since v225 does this for selected gpio sources as well (commit). A bit of digging shows that:

  • The udev rules add the "power-switch" tag to selected devices. Udev does not use these tags itself, but other programs can.
  • The "power-switch" udev tag causes systemd-logind to open the event source as a possible source of power events.
  • When a KEY_POWER keypress is received, systemd-logind initiates a shutdown. It also handles some suspend-related keys.
  • Since last week, 70-power-switch.rules tags all event sources as "power-switch" (commit) and lets systemd-logind figure out which could potentially fire KEY_POWER events (commit) and monitor only those (though this change does not really change things for this usecase).

Linux gpio-keys driver

So, how to let a GPIO pin generate a KEY_POWER event? The first logind commit linked above has an example device tree snippet that sets up the kernel gpio-keys driver which does exactly that: Generate key events based on gpio changes.

Normally, the devicetree is fixed when compiling the kernel, but using a devicetree overlay (see the Rpi device tree docs) we can insert extra configuration at boot time. I found an example that ties multiple buttons to GPIO pins and configures them as a tiny keyboard. Then I found an example in the the gpio-button package that does the same with a single button and also shows how to configure a pulldown on the pin.

I've created a devicetree overlay based on that example, that sets up the gpio-keys driver with a single GPIO to trigger a KEY_POWER event. The overlay is just a text file and I added elaborate comments, so look inside if you are curious. Instructions on installing this overlay are at the end of this post.

Waking up

With the above, you get a shutdown button: Press it to initiate a clean shutdown procedure, after which the raspberry pi will turn off (note that it does not completely cut power, that would require additional hardware. Instead, the system goes into a very deep sleep configuration while still drawing some current).

After shutdown, starting the system back up means removing and reinserting the USB cable to momentarily cut power. This begs the question: Can you not start the system again through one of the GPIO pins?

I found one post that pointed out the RUN pin, present on Rpi 2 and up, which can be shorted to GND to reset the CPU (without clean shutdown!), but can also power the system up. However, having two buttons, along with the risk of accidentally pressing reset instead of shutdown did not seem appealing to me.

After more digging I found a forum post saying that shorting GPIO3 to GND wakes up the Rpi after shutdown. It also says GPIO1 should work, but I do not think any Rpi makes that pin available, so I haven't tested that. Official docs on this feature seem to be lacking, I found one note on the wiki saying bootcode.bin (version 12/04/2012 and later) puts the system in sleep mode, and starts the system when GPIO3 goes low.

Since GPIO3 is also a normal IO pin, this is perfect: By configuring the shutdown handling on GPIO3, the same button can be used to shutdown and wake up. Note that GPIO3 is also an I²C pin, so if you need to also connect I²C-devices, another pin must be used for shutdown and you cannot use the wakeup feature. Also note that GPIO3 (and GPIO2) have external pullups at least on the Rpi Zero.

Setting this up

  1. Get and build the DT overlay. If a file /boot/overlays/gpio-shutdown.dtbo is already available on your system, you can skip this step.

    Download the devicetree overlay file. The easiest is to run wget on the raspberry pi itself:

    $ wget

    Compile the devicetreefile:

    $ dtc -@ -I dts -O dtb -o gpio-shutdown.dtbo gpio-shutdown-overlay.dts

    Ignore any "Warning (unit_address_vs_reg): Node /fragment@0 has a unit name, but no reg property" messages you might get.

    Copy the compiled file to /boot/overlays where the loader can find it:

    $ sudo cp gpio-shutdown.dtbo /boot/overlays/
  2. Enable the overlay by adding a line to /boot/config.txt:


    If you need to use a different gpio, or different settings, see the dts file for available options and the Rpi devicetree docs for setting them.

  3. If running systemd older than v225 (check with systemd --version), create a file called /etc/udev/rules.d/99-gpio-power.rules containing these lines

    ACTION!="REMOVE", SUBSYSTEM=="input", KERNEL=="event*", SUBSYSTEMS=="platform", \
        DRIVERS=="gpio-keys", ATTRS{keys}=="116", TAG+="power-switch"

    On systemd v225 and above (Raspbian stretch version 2017.08.16 and upwards) this should not be needed, but I have not tested this.

  4. Reboot

Then, if you connect a pushbutton to GPIO3 and GND (pin 5 and 6 on the 40-pin header), you can let your raspberry shutdown and startup using this button.

All this was tested on a Rpi Zero W, Rpi B and a Rpi B+.

This overlay was merged into the official kernel repository, so in the future step 1 above should no longer be needed. It is included in the 1.20170811-1 kernel release, which will hopefull be included in Raspbian images soon (but the 2017.08.16 image does not include it yet).

When the new kernel is included in Raspbian, it would only leave a simple modification to /boot/config.txt to set this up :-D

Gamemaster wrote at 2017-07-24 09:33

Sadly this did not work for me. However, the solution at 8bitjunkie did if you want to compare notes.

Matthijs wrote at 2017-07-24 09:41

Hm, pity. I wonder what part is failing exactly (the link you gave has a completely different software approach, so hard to compare). If you want to help out debugging this, drop me an e-mail at and I'll be happy to suggest some tests to see what is failing exactly.

Windy City Jess wrote at 2017-08-03 06:13

Matthijs, I've tried many solutions with no luck. It works for me now that I followed your directions. You were right on all accounts, the RPi device-tree page you linked above is very helpful and systemd 215 did require 99-gpio-power.rules. Initially skipped over this step because I assumed it was optional. I should learn to read and follow directions :) Thank you for sharing this. - Jess

Blistex77 wrote at 2017-08-18 10:55

I've been looking for a simple solution for a while, I'm quite a noob, but it looks simple enough to achieve, thanks a lot! I have only one question, to be sure, I would only need a jumber cable and a power button (one in this kit for example

Matthijs Kooijman wrote at 2017-08-18 11:31

That's correct, you basically only need something that can momentarily connect two pins together. A screwdriver would suffice, but two jumperwires (or one cut in half soldered to the button) and a pushbutton would be more obvious :-)

Blistex7 wrote at 2017-08-18 11:59

Thanks for the fast reply, I'll try this asap!

John w wrote at 2017-08-22 19:04

does the dt overlay collide with i2c? I want both. On this project ill probably resort to an on off ic but not having i2c if you want this functionality with out ic is kinda a deal breaker. pi doesnt have a lot of io to begin with

Matthijs Kooijman wrote at 2017-08-22 19:21

Yeah, this overlay collides with I²C. I haven't tried myself, but got a report from someone that ran into problems. Note that you can easily switch to another pin for the shutdown feature, you'll just lose the button-based power-on (but if you really need it, you could wire a separate reset button to the RUN pin, which should also power on after power off).

Dom C wrote at 2017-08-24 00:22

What would happen if your switch was latching?

I have s nice LED lit one around that would be perfect for this but feel it might not work in this case?

Matthijs Kooijman wrote at 2017-08-24 00:54

I believe that when the pi is powered on, it behaves like a keyboard key, so it sends a keydown when the switch is closed and a keyup when the switch is opened. I'm not sure what systemd responds to, probably both. When the pi is powered down, I believe it only responds to a falling edge, so the switch closing.

This means that if you use a latching switch, things might just work - close the switch to power up, open the switch to generate a keyup and power down. It could be that a keyup is not sufficient, or that on startup a keydown is generated. In that case, it might be needed to set the active_low flag to 0, so you get a keydown on a rising edge instead. One risk of this approach is that the pi state and the button state might become out of sync, but that should be easily fixable by just toggling the button once or twice.

Dom C wrote at 2017-08-27 19:06

I bit the bullet an bought a push-to-make button anyway. All working great until i tried to add a seperate GPIO pin for a Power LED.

I added; dtoverlay=gpio-poweroff,gpiopin=6,active_low

But this seemed to cause some kind of clash? No longer did my button (using GPIO3 as suggested) all me to power back on (although shutdown continued to work fine)

Any thoughts?

Matthijs Kooijman wrote at 2017-08-29 18:46

@Dom, I'm not entirely sure what you're saying, but if you change the pin number in the overlay, that only influences the shutdown, the power-on is hardcoded in the bootloader (or something like that). For you it seems to be reversed, which I can't quite explain?

Also, note that active_low should normally get a value of 0 or 1, and it defaults to 1 (active low). To make it active high, use active_low=0.

uzanto wrote at 2017-08-30 12:25

Thank you mate!

Working in my RPi without the udev rule.

Raspberry Pi 3 B

Raspbian Stretch with kernel 4.9.41-v7+ SystemD 232

I need to make the directory overlays inside /boot/

Simon wrote at 2017-09-01 21:42

Pi ZeroW with Stretch. This seems to shutdown kind-of (ssh and other servers become unreachable, even usb otg serial console dies) but it still responds to ping, then won't start up again. Any ideas?

Matthijs Kooijman wrote at 2017-09-04 10:11

@Simon, interesting, seems like the shutdown fails somehow. Did you install extra packages that might interfere? Does it work with a clean stretch system? Does a normal shutdown from the commandline (shutdown -h now and systemctl poweroff) work? Anything interesting in the syslogs? Perhaps redirect the primary console to serial (with the console= kernel commandline option, not sure how to set that or if that works with an USB-otg serial port) and see if anything shows there?

Arjen wrote at 2017-09-23 17:43

Is there a way to insert a delayed reaction to the button pressed? I do not want it to shut down if I accidentally hit the button, it should be pressed for e.g. 1 sec?

Matthijs Kooijman wrote at 2017-09-25 12:05

Looking at the gpio-keys documentation, it does not support a "delayed press":

However, IIRC the gpio-keys driver does generate proper keydown and keyup events, so systemd could, in theory, detect the length of a press and require a minimum length before shutting down the system. I do not think systemd supports this now, but perhaps the maintainers would be willing to add it. You could see if there's a feature request for this already at and, if not, add one? If you find or add one, please post the link here!

John Shearing wrote at 2017-09-26 08:07

Worked perfectly! Thank you so much for the solution, the explanation, and for taking the time to link this page in other forums so that it could be found.

John Shearing wrote at 2017-09-28 22:34

I have three questions please:

  1. Pin 5 is quite sensitive. Just attaching or removing a small wire which is not connected to anything else causes the pi to shut down or start up. Can something be done to make this pin less sensitive?

  2. I am powering the pi with a battery and charging it a circuit which is capable of completely removing power from the pi after your shutdown process has completed it's work. A URL to the charging device is as follows: In order to tell the charger to kill power I need a pin on the pi which goes from 3.5 volt to 0 volts after your shutdown process has run. Is there a way to make a pin behave this way? If so, please tell me how.

  3. I am running the GDM3 display manager because it is the only login screen I have found which incorporates a virtual onscreen keyboard. If I have started the pi but have not yet logged in and I short pins 5 and 6 together then your process will blank the screen but will not shut down the pi. Shorting the wires again immediately returns the login screen. My question is: Is it safe to completely remove power when the pi is in this condition or must I first find a way to execute the shut down command.

Any further guidance you can give would be greatly appreciated. Thanks again for all your help. John

Matthijs Kooijman wrote at 2017-10-01 10:31

@John, I'll try answering your questions.

  1. The pin should have a pullup resistor (internal to the CPU and external because of I²C) that prevents it from being super-sensitive. Perhaps your using a raspberry pi without an external pullup? In any case, using a smaller pullup (which needs more current to drop 3.3V) could help there. Typical values are between 1k and 100k. You could try 1k (or even smaller, but the smaller the pullup, the more current flows when you press the button).

  2. I believe there is a "gpio-poweroff" kernel module and overlay for this usecase, but I haven't looked at it at all.

  3. The handling of the power button normally happens in systemd. I suspect that GDM3 "inhibits" the power key handling in systemd and then handles it in a different way itself. Perhaps you could configure GDM3 to not handle the power keys at all, or to handle the power key by doing a shutdown just like systemd normally does?

John Shearing wrote at 2017-10-07 02:41

Hi Matthijs Sorry for the late response. It has taken me some time to absorb all the information you have given us. Truly everything we need to work with overlays is referenced in your post.

Thanks for explaining how to use a pull up resistor.

The gpio-poweroff overlay worked perfectly. It was already on the pi in the overlays directory. I called it in config.txt right after calling your overlay.


Directions for using all the overlays found in the overlays directory are in the README which is found in the same overlays directory.

For some reason the gpio-poweroff overlay interferes with restart functionality of pin 5 but that doesn't matter for what I am doing. It still halts the pi just fine.

I am now researching the problem where the overlays don't seem to load until after GDM3 login or as you suggest, the overlay behavior is overridden by GDM at the systemd level. I will report back if a solution is found.

Thanks again for all your help.

dqbydt wrote at 2017-10-10 01:42

This mechanism works well on a CM3+CMIO3. One question: you mentioned in an earlier comment "the power-on is hardcoded in the bootloader". Do you have any more information regarding this? Is there no way to configure the GPIO pin for powerup? I cannot use GPIO1 or GPIO3 in my application because both are irrevocably tied to other functions. I am able to powerup by grounding the RUN pin, but would be nice to have shutdown as well as powerup on the same GPIO.

Matthijs Kooijman wrote at 2017-10-10 08:55

> Do you have any more information regarding this? Is there no way to configure the GPIO pin for powerup?

All info I have is linked in the post above. AFAIK, the wakeup pin is hardcoded in bootcode.bin, for which no sources are available. However, re-reading the forum post linked above, it says "I've now added a way to wake the board through GPIO.", which suggests that whoever wrote that does have access to the bootloader source. Perhaps you could try asking him or her (or other official rpi engineers) about this? If you find out anything, please report back here!

dqbydt wrote at 2017-10-11 17:57

I did ask on RPi forums, and this is the reply from the engineer (dom) who originally added the wakeup support:

"The low power mode following "halt" is achieved by not powering on a number of power domains/clocks. That means no access to sdcard and hence no support for configuration files to choose wakeup pin."

So I guess you are restricted to using GPIO1/3 for wakeup.

JRoth wrote at 2017-12-02 19:54

Thank you for making this available, it works much better than the python solutions I've tried.

Comments are closed for this story.

25 comments -:- permalink -:- 10:51
Retraining the Spamassassin Bayes filter with recent messages

Tags: Software, Spam, Mail, Linux

On my mailserver, I'm using Spamassassin with a Bayes filter to detect spam. Such a filter needs to be trained with samples of spam and ham (non-spam) messages to let it learn what spam and ham looks like, but it also needs to be retrained when the spam or ham changes over time. I have some automatic training set up, but since a while I've seen the bayes filter being completely wrong (showing a confident ham score for something that is very clearly spam), so I decided to retrain the filter from scratch, using the spam and ham messages I collected over the last time (I don't really throw away any e-mail).

Since training with all my e-mail is not productive (more than 5,000 messages aren't really helpful AFAIU, and training with old messages is not representative for current messages), I decided to just take all of my e-mail and take the last 2,000 spam and ham messages and train with that. My spam is neatly collected in 2 mailboxes (Spam for obvious spam and ProbablySpam for messages that need an occasional review to find false positives), but my ham is sorted out in dozens of different mailboxes. Hence, I needed some find magic to get a list of the most recent spam and ham messages. So, I built these commands:

# find Spam ProbablySpam -type f \( -path '*/cur/*' -o -path '*/new/*' \) -printf "%T@ %p\n"
  | sort -n | cut -d' ' -f 2 | tail -n 2000 > spam
# find . -type d \( -path ./Spam -o -path ./ProbablySpam -o -path ./Bulk -o -path ./Sent \) -prune -o \
         -type f \( -path '*/cur/*' -o -path '*/new/*' \) -printf "%T@ %p\n" \
  | sort -n | cut -d' ' -f 2 | tail -n 2000 > ham
# sa-learn --progress --spam -f spam
# sa-learn --progress --ham -f ham

After retraining with recent spam, the results were a lot better, so I'm not longer spending time every day deleting a couple dozens spam e-mails :-D

0 comments -:- permalink -:- 21:03
Using Xctu through an Arduino shield

Tags: XBee, Arduino, Electronics, Nerd

XBee modules are a range of wireless modules built by Digi, and are typically used to add wireless connectivity to Arduino or other microcontroller based projects. To configure these modules and update their firmware, you can use the XCTU configuration utility. This utility uses a serial port to talk to the XBee module, so you will need some way to connect to the XBee module to a serial port on your computer (using a USB "TTL" serial port, a "real" RS232 port has too high voltage).

The easiest way is to use a dedicated board, like the SparkFun Explorer USB:

SparkFun Explorer USB

However, if you already have an Arduino and an XBee shield for it, you might want to use those to connect XCTU to your XBee module. In theory, this should be a matter of re-arranging some wires, but in practice I've run into some problems attempting this (depending on the hardware used).

In this post, I'll show a few ways to do this using an Arduino and a shield, and explain some of the problems you might run into.

First, let's see what connections we really need. When wiring things up using an XBee explorer, these are the connections made (leaving out stuff like the reset button, leds, and power and ground connections):

SparkFun Explorer simplified schematic

Here, the two data lines (TX/RX/DIN/DOUT) are the most important, those are needed for configuration and normal firmware updates. The other handshaking/flow control signals can prevent overflowing buffers at high speeds, but that is not normally a problem. They might also be needed to recover from a failed firmware upgrade, to force the XBee module into bootloader mode.

So, how can you use an Arduino and XBee shield to achieve these connections? Below, three variants are explored, each with different requirements, advantages and disadvantages.

Bypassing the Arduino microcontroller

Most Arduinos contain a single microcontroller (MCU), and have a USB-to-serial converter integrated on the board as well. This USB-to-serial converter is connected to the microcontroller, so it can be used to upload programs to the Arduino, and send debug output using Serial.println(). The connections on the Arduino Uno board are like this:

Arduino simplified schematic

The resistors you see here are protection resistors, to prevent a short circuit when you connect another serial device to the Arduino's RX pin, but at the same time the USB-to-serial device also enables its TX pin. As you'll see later, these resistors might cause issues as well.

Now, to connect the XBee to XCTU, you can connect it to the USB-to-serial converter. To do so, you should connect the DOUT pin to the Arduino's TX pin, and the DIN pin to the Arduino's RX pin. This seems wrong, but remember that you want to talk to the USB-to-serial chip, not the Arduino. This would look like this (voltage converters not shown):

Bypassing Arduino

How to obtain this configuration depends on the shield used. The official Arduino Wireless SD shield, as well as some others, have a switch that can enable this wiring (in the other position, the connections are reversed, making the XBee module talk to the microcontroller instead). On some shields, like the SparkFun XBee shield, there is switch that can connect the the XBee module to pin 2 and 3 on the Arduino. Then, by connecting pin 0 to 3 and 1 to 2 on the shield's pin header, you get the above configuration. Yet other shields, like the one from ITEAD, simply have a long row of jumpers that allow connection DIN and DOUT to any I/O pin you want.

Interfering microcontroller

Note that the microcontroller is also still connected, since the Arduino board has no way to disconnect it. To prevent the microcontroller from interfering with the XBee communications, you will have to actually remove the microcontroller from the socket. This is not possible with all Arduino boards, though. If you do this, the auto-reset stuff below does not apply, though the pullup problems do.

You might think (I did) that you can also just upload an empty sketch to the Arduino, which keeps its serial port disabled. However, when XCTU opens up the serial port, the Arduino resets and the bootloader runs for a second or so, enabling the the serial port during that time. Because XCTU sends some data, the bootloader keeps running for longer, making the communication between XCTU and the XBee module fail.

Disabling auto-reset

This can be fixed by disabling auto-reset, but the only reliable way I've found to do this is to cut the "RESET EN" solder jumper. Since you'll have to re-solder the jumper again to enable it (needed for uploading new sketches), this seems a hassle. A lot of people online mention fiddling with a specific value resistor or capacitor to disable auto-reset, but that didn't work for me (I think the point is making the reset pulse so short it goes undetected, which is of course a bit shaky in any case). Connecting RESET directly to the 5V pin worked for me, but this creates a very brief short circuit whenever it would otherwise auto-reset (which might damage the usb-to-serial chip, though I'm not sure if this is a real risk).

Instead of disabling auto-reset, you could also try to keep the Arduino in reset by connecting RESET to GND. Again, this works, but has the same brief short circuit described above.

Problems with pullups

Even if you solve these problems, you might run into further problems, depending on the XBee shield used. Since Arduinos typically run at 5V and XBee modules at 3.3V, shields need voltage converters (which were hidden in the previous image). For example, the SparkFun XBee shield uses a transistor and two pullup resistors for this converter. This results in the following (only the DIN line is shown, DOUT is identical):

Pullups causing trouble

Since the pullups used are 1kΩ, just like the protection resistors in the Arduino board, communication will break. When the USB-to-serial converter pulls its TX line low, a resistor divider is created, resulting in 2.5V on the transistor instead of 0V, preventing it from switching on. A similar problem can occur when a resistor divider is used for voltage conversion, like the ITEAD shield does, though that particular shield uses big enough resistors (10kΩ) to not cause any problems.

So, this bypass approach is only limited in use. If you use it, you can do both configuration as well as firmware updates in XCTU. Recovering from a failed firmware upgrade is sometimes possible, but since no handshaking connections are available, this does not always work (in particular, it only works if the bootloader detects that the firmware update has failed, so the bootloader stays active all the time).

Forwarding through the Arduino microcontroller

Instead of bypassing the Arduino microcontroller, you can also forward data through the Arduino microcontroller. One one side, the Arduino talks to the computer running XCTU, on the other side the Arduino talks to the XBee module. A simple sketch inside the Arduino takes care of forwarding the data.

Since most Arduinos only have one hardware serial port, connected to the USB-to-serial converter, you will need to use a software serial port to talk to the XBee. I recommend using the AltSoftSerial library for this, which is significantly more reliable (it can receive while sending, for example) than the Arduino-supplied SoftwareSerial library.

The wiring would look something like this (note that AltSoftSerial dictates the use of pin 8 and 9 on an Arduino Uno):

Forwarding through Arduino

Achieving this wiring can be a bit tricky. On shields like the ITEAD one, this is a matter of placing the jumpers correctly. On shields like the SparkFun one, this means setting the switch so the XBee connects to pin 2 and 3, and then connecting pin 2 to 8 and 3 to 9. On shields like the Arduino Wireless SD shield, this wiring is impossible without cutting traces.

Disabling auto-reset

When using this approach, auto-reset again causes problems. When XCTU opens the serial port, the Arduino will reset and XCTU ends up talking to the bootloader instead of your sketch. Disabling auto-reset, as described above, helps here.

Fixed baudrate

When you use this approach, the forwarding sketch will need to configure a fixed baudrate for both serial connections. Now, since configuration, as well as triggering a firmware upload, typically happens at 9600 baud and the actual firmware upload at 57600 or 115200 (depending on the XBee board used), updating the firmware is not possible with this approach — you would have to (very quickly) switch the baudrates somewhere during the first part of the firmware update. In theory, you could first change the configuration baudrate to 57600 or 115200, so that the same baudrate can be used for both, but I haven't tried this approach.

Forwarding through the Arduino Leonardo

When using the Arduino Leonardo (or another ATmega 32u4-based board), this problem can be solved in a much better way. The Leonardo does not have a separate USB-to-serial converter, but instead its microcontroller supports USB directly, so it can expose a virtual serial port on the computer itself. This is useful, because:

  1. This leaves the single hardware serial port on the Leonardo available to talk to the XBee module.
  2. The auto-reset on the Leonardo works differently, so XCTU will never accidentally trigger auto-reset.
  3. The sketch has access to the baudrate configured by XCTU, as well as the handshaking signals, allowing both firmware updates and full firmware recovery.

Wiring is also simplified, because all shields trivially support connecting the XBee module to the Arduino hardware serial port on the RX and TX pins. This is what the wiring looks like:

Forwarding through Arduino Leonardo

Note that this additionally connects the XBee handshaking signals to Arduino I/O pins, to support firmware recovery. If this is easy depends on the shield again - typically you will need to solder some wires, though I believe the ITEAD shield already has these connections in place. You can leave them out if you just want to do configuration and firmware uploads.

The sketch that implements this isn't so trivial, so I've made it available for download. This sketch uses some recently introduced changes in the virtual serial port to get access the baudrate and handshaking signals, so this sketch will need the Arduino AVR core 1.6.9 (expected to be shipped with Arduino IDE 1.6.6). Neither of these is released yet, so for now you can use the Arduino hourly build instead.


So, it turns out that using an Arduino and XBee shield for talking to your XBee module from XCTU is not quite as easy as you would expect. Using an Arduino Leonardo works perfectly, but other approaches are either cumbersome, or potentially damaging for the hardware... If anyone is aware of better approaches, please let me know!

0 comments -:- permalink -:- 16:23
Interrupts, sleeping and race conditions on Arduino

Tags: Nerd, Hardware, Electronics, Arduino, Sleeping, Powersaving

Arduino Community Logo

My book about Arduino and XBee includes a chapter on battery power and sleeping. When I originally wrote it, it ended up over twice the number of pages originally planned for it, so I had to severely cut down the content. Among the content removed, was a large section talking about interrupts, sleeping and race conditions. Since I am not aware of any other online sources that cover this subject as thoroughly, I decided to publish this content as a blogpost separately, which is what you're looking at now.

In this blogpost, I will first explain interrupts and race conditions using a number of examples. Then sleeping is added into the mix, which again results in some interesting race conditions. All these examples have been written for Arduino boards using the AVR architecture, but the general concepts apply equally well to other platforms.

The basics of interrupts and sleeping on AVR are not covered in detail here. If you have no experience with this, I recommend these excellent articles on interrupts and on sleeping by Nick Gammon, which cover interrupts, sleeping and other powersaving in a lot of detail.

In this post, I will show relevant snippets of the code, but omit trivial things like constant definitions or pinmode settings. The full example sketches can be downloaded as a tarball and each will be separately linked below as well.

Blinky light using interrupts

This first example will explore the use of interrupts, starting without sleeping, then sleeping will be added. The sketch will light up the internal led whenever a button is pressed and keep it lit until the button is not pressed for four seconds. Button presses will be detected using an external interrupt, using the Arduino attachInterrupt() function.

Turning the LED on

When an interrupt happens, an ISR function will be called. Note that when using attachInterrupt(), you are not defining a real ISR, but just a normal function that gets called by the ISR that is hidden inside the Arduino code. This "ISR" looks like this:

// Time of the last buttonpress
volatile uint32_t last_press = 0;

// On a buttonpress: turn on led and record time
void buttonPress() {
  digitalWrite(LED_BUILTIN, HIGH);
  last_press = millis();

Pressing the button turns on the LED, and remembers the timestamp of the buttonpress. If the LED is already on, the LED will be unchanged, but the timestamp will be updated.

To keep this timestamp, a global variable is defined, so it can be accessed both from the interrupt handler and the loop. Note that it is declared with the keyword volatile. This keyword tells the compiler that the variable is used from inside an interrupt handler. Formally, it is a bit more complicated than that, but in most cases it is ok to remember to use volatile on all variables that are written to inside an interrupt handler and read or written outside an interrupt handler. This keyword tells the compiler that the variable can change at any time, and that the compiler should not optimize away access to this variable.

To make sure that this function is called when the button is pressed, it has to be registered with the Arduino code:

void setup () {
  // Set up button
  attachInterrupt(BUTTON_INT, buttonPress, FALLING);

  // Set up led

This sets up the input and output pins and registers the buttonPress() interrupt handler. It should get called on every falling edge (so when the button is pressed).


Note that this example completely ignores switch bouncing (which is the effect that when you push or release a switch, it will very swiftly connect and disconnect a few times, causing the interrupt handler to trigger multiple times). Bouncing is not a problem for this example, but check out this article for more info on connecting switches, including some strategies for debouncing.

Turning the LED off

So, now you have a way to turn the light on, but you also need to turn it off when a timeout has passed. This is handled by the loop() function:

void loop () {
  if (millis() - last_press >= TIMEOUT)
    digitalWrite(LED_BUILTIN, LOW);

Here, (millis() - last_press) is the time since the last button press. Whenever that time becomes larger that TIMEOUT, the led is turned off (and, until the button is pressed again, it will continue being turned off every loop, which is not terribly useful, but won't hurt either).

Note that this particular way of handling things correctly handles millis() overflow. For more info, see this article.

If you upload the sketch to your Arduino board, you will have a button-controlled blinky light with timeout. Perfect! Or is there perhaps still a problem?

Race conditions

Perhaps, while playing with your brand new blinky toy, you noticed it did not always work as expected. If not, go ahead and press the button repeatedly. As long as you press it at least once every four seconds, the led should always remain on, right? If you keep pressing the button (as fast as you want), you will see that sometimes the led actually turns off anyway. It might need a couple dozen of presses, but it should happen eventually.

How is this possible? To understand what is going on, you will have to understand that the AVR microcontroller is an 8-bit microcontroller. This means that most of its operations, and in particular accessing memory, happen one byte (8 bits) at a time.

Note that the last_press variable is a uint32_t variable, meaning it is 32 bits, or 4 bytes long. So when the loop() function needs to read it, each of these bytes are read from memory, one by one, using separate instructions.

Now consider what happens when the pin interrupt triggers in the middle of this operation? loop() will have fetched some of the bytes but then the interrupt handler will change the variable in memory, after which the loop() continues to fetch the remaining bytes from memory. The result is that half of the bytes come from the old value, while half of the bytes come from the new value. This will likely cause the comparison to completely mess up and return true even when the timeout has not expired yet.

If this seems rather unlikely: you have seen it happening, perhaps even multiple times if you kept going for a while. Since the microcontroller is not doing a whole lot except for checking the last_press variable over and over again, the chance of the interrupt triggering at the exact right time is actually fairly significant.

What you are seeing here is what is commonly called a race condition. Generally speaking, a race condition is present when two things need to happen in a certain order but it depends on chance whether they will actually happen in the right or wrong order. Typically, when a race condition is present, the events usually happen in the correct order, and only rarely the incorrect order occurs. This often makes race conditions particularly hard to reproduce, with problems occurring occasionally in your software in production, but never on the developer's system where they could be diagnosed. Because of this, recognizing race conditions early is a big win. Whether you are dealing with interrupts in a microcontroller, multiple threads or processes in a bigger operating system, or true concurrency with multiple processors or cores, wherever there are multiple concurrent threads of execution, race conditions will be lurking around the corner.

In this case, the correct ordering of events is that the interrupt should be handled either before or after all four bytes of last_press are loaded, but not in between the loading of the bytes. Another term often used for this is to say that last_press must be loaded atomically, meaning it should not be possible to be interrupted halfway.

If you look closely, you might find there is a second race condition. Consider what happens what happens when the interrupt triggers after the last_press variable was loaded, but before the led is disabled?

In this case, the led will be turned on by the ISR, but it is immediately disabled again by loop() (which had already decided it would turn off the led). So instead of staying on for 4 more seconds, the led stays off. Triggering this bug requires pressing the button at the exact moment the light is about to turn off, so it is very unlikely that you will trigger this bug in your testing. But if you would design a device containing this code and produced it a million times, some of your users will probably see the bug. It is common to say that the window for triggering this bug is very small.

The most common way to fix this is fairly simple: disable interrupts around code that needs to be executed atomically. This ensures that an interrupt cannot occur during the code, guaranteeing correct ordering. If any interrupt triggers while interrupts are disabled, the CPU will queue the interrupt (by setting an interrupt flag bit in a register) and as soon as interrupts are enabled again, the interrupt handler will run.

Usually, you should try to only disable interrupts for a very short time. The longer interrupts are disabled, the longer any queued interrupts will have to wait, which can cause problems with timing-sensitive applications (just like interrupt handlers themselves must be short).

So, what does this mean for the code? Just disable interrupts before the if and re-enable them afterward. This ensures that loading the last_press value, but also the decision and turning off the led now happen atomically, forbidding the interrupt to trigger halfway. This looks like this:

void loop () {
  if (millis() - last_press >= TIMEOUT)
    digitalWrite(LED_BUILTIN, LOW);

This uses the noInterrupts() and interrupts() functions defined by Arduino to disable and re-enable interrupts globally (meaning all interrupts are disabled). You might also encounter cli() (clear interrupt bit) and sei() (set interrupt bit), which are the AVR-specific versions of the same functions. Using the Arduino versions makes it easier to port the code to other architectures too.

With this change applied to the sketch, you should be able to keep punching the button over and over again, with the led staying on (until you stop pressing for four seconds, of course).


Now that you have an interrupt controlled button working, time to add sleeping. One of the reasons to use interrupts, is that (only) an interrupt can wake an Arduino from its sleep. This means that with the code shown above, once the LED is off and the Arduino is waiting for the next button press, it can just go to sleep, knowing it will be woken up when a button is pressed:

void loop () {
  if (millis() - last_press >= TIMEOUT) {
    digitalWrite(LED_BUILTIN, LOW);

This is just the previous loop() function, with the doSleep() call added after the led is turned off (the doSleep() function is shown below). Since a button press wakes up the microcontroller, it will end up waiting for a button press in slumber and only resuming with the code after doSleep() when a button was pressed.

Note that doSleep() is called with interrupts disabled. You might think it would be good to re-enable interrupts after turning off the LED, to keep them disabled as short as possible. However, this would introduce another race condition: Consider what would happen if the button interrupt would trigger after turning off the led, but before going to sleep? In this case, the interrupt handler would turn on the led, and then the microcontroller goes to sleep. During sleep the loop() function will not run to detect that four seconds have passed, so the led will stay on indefinitely (until pressing the button wakes up the microcontroller again).

By keeping interrupts disabled when calling doSleep(), this is avoided. However, the interrupts cannot remain disabled when actually going to sleep, since then the microcontroller can never wake up again. So they should be re-enabled just before going to sleep:

void doSleep() {
  set_sleep_mode (SLEEP_MODE_PWR_DOWN);

  sleep_cpu ();

But what if a button press happens between disabling interrupts and re-enabling them again just before sleeping? As interrupts are disabled, the interrupt handler cannot run, but a flag will be set and the ISR will run after interrupts are enabled again. Now, the AVR architecture guarantees that after interrupts are enabled, at least one instruction runs uninterrupted. In this case, this means that the sleep_cpu() function (which translates to a single sleep instruction) always runs. If an interrupt is pending, it will be processed after the sleep instruction, causing the microcontroller to wake up immediately again (which is not terribly efficient, but it is correct).

So, with this sketch, the Arduino is sleeping while the LED is off, significantly reducing power usage. However, when waiting for these 4 seconds to pass, the Arduino is still running and consuming power. To also sleep while waiting, you will need some way to wake up after 4 seconds have passed.

Timed wakeups

With what you have seen so far, you could add an external timer that pulls a line high or low after some time. If you connect this line to an Arduino interrupt pin, you can have the Arduino wake up at the right moment. In fact, this approach is sometimes used combined with a real-time clock (RTC) module, which is particularly good at keeping accurate time and tracking long timeouts.

Fortunately, the AVR microcontrollers also feature a number of internal timers that can be used for this purpose. The most power efficient timer is the watchdog timer. Its original purpose is to automatically reset the microcontroller when it is locked up, by keeping a counter and resetting the system if the software does not regularly reset this counter.

However, the watchdog module has a second mode, where it does not reset the entire microcontroller, but just fires an interrupt. Since the watchdog timer runs on a private 128kHz oscillator, it can also run while in power-down mode (where all other clocks are disabled) and its interrupt can cause a wakeup. In addition to being low-power, this oscillator is not very accurate and cannot handle arbitrary intervals. If you need a more precise timing source, you should look at using power-save mode and run timer2 in asynchronous mode, using a secondary crystal.

To use the watchdog timer for the four-second timeout, the buttonPress() interrupt handler should be modified as follows:

void buttonPress() {
  digitalWrite(LED_BUILTIN, HIGH);

As before, this turns on the led. However, instead of keeping a last_press timestamp, this resets the watchdog timer (restarting its counter if it was already counting) and it enables the watchdog and configures it to count up to four seconds in case it was not running yet.

Note that this uses the enable_wdt_interrupt() function to enable the watchdog timer interrupt, but not the watchdog timer itself. Unfortunately, avr-libc does not provide a function for this, so this uses a custom wdt_interrupt.h file that you need to put alongside the sketch.

Now, when the watchdog timer expires, an interrupt is triggered. This wakes up the microcontroller and then runs the WDT_vect interrupt handler, which is defined as follows:

ISR(WDT_vect) {
  digitalWrite(LED_BUILTIN, LOW);

When the timeout happens, the led is turned off, and the watchdog is disabled again (it will be re-enabled when you press the button).

Since both of the interrupt handlers above already take care of all behavior needed, there is nothing left for the loop() function to do other than to sleep:

void loop () {

Now with this sketch all the race conditions you fixed earlier can no longer occur now, since the loop() function no longer makes any decisions about whether to sleep, and the two interrupt handlers cannot interrupt each other (so each will run atomically already).

So, hopefully after reading this post, you will have a better idea of the challenges and race conditions involved when making an Arduino sleep. If you have further questions or remarks, feel free to leave them below!

0 comments -:- permalink -:- 17:34
All chapters done!

Tags: XBee, Book, Arduino, Electronics, Personal, Nerd

Yay! Last night (yes really night, it was 4:30 AM) I finalized and submitted the last chapter of my book. There's still a few details left, but the text of the book is out of my hands now, and the great folks at Packt publishing are now doing their magic with correcting my English, testing my code, layouting the book, and so on.

Great to have the pressure of these (repeatedly postponed) deadlines off my back, so I can finally take some time to do some long-due other stuff. Like cleaning up my desk:

Mess on my desk

Constantine wrote at 2016-01-31 16:54

Hello, yesterday I just bought your book and i have some questions about chapter 5.

Matthijs Kooijman wrote at 2016-02-01 12:21

Constantine, feel free to drop me a note at with your questions!

Comments are closed for this story.

2 comments -:- permalink -:- 16:27
Writing "Building Wireless Sensor Networks Using Arduino"

Tags: XBee, Book, Arduino, Electronics, Personal, Nerd

Building Wireless Sensor Networks Using Arduino

Past summer, I've been writing a book, which is now nearing completion. A few months ago I was approached by Packt, a publisher of technical books, which want to add a book about sensor networks using Arduino and XBee modules to their lineup, and asked me to write it. The result will be a book titled "Building Wireless Sensor Networks Using Arduino".

I've been working on the book since then and am now busy with the last revisions on the text and examples. Turns out that writing a book is a lot more work than I had anticipated, though that might be partly due to my own perfectionism. In any case, I'll be happy when I'm done in a few weeks and can spend some time again on all the stuff I've been postponing in the last few months :-)

0 comments -:- permalink -:- 12:12
Sailing trip to the Isle of Wight (Sponsored by Murphy)

Tags: Vacation, Sailing, AIS, Murphy

Atlantis under sail

I originally wrote this, just under a year ago when I returned from this sailing trip. I found this draft post again a few weeks ago, being nearly finished. Since I'm about to leave on a new sailing trip next week (to France this time, if the wind permits), I thought it would be nice to finish this post before then. Even though it's a year old, the story is bizarre enough to publish it still :-)

I've put up photographs taken by me and Danny van den Akker up on Flickr.

Last year in May, I went on a 11 day sailing trip to the Isle of Wight an island off the southern coast of England. Together with my father, brother and two others on my father's boat as well as 10 other people on two more boats, we started out on this trip from Harderwijk on Thursday, navigating to the port of IJmuiden and going out on the north sea from there on Friday morning.

Initially, we had favourable wind and made good speed up to Dover at noon on Saturday. From then on, it seemed Murphy had climbed aboard our ship and we were in for our fair share of setbacks, partly balanced by some insane luck as well. In any case, here's a short timeline of the bad and lucky stuff:

  • :-( Just around Beachy Head, our engine stopped running. We quickly restarted it and carried on, but shortly after the engine stopped again. It didn't want to start again either. In case you're wondering
    • not having a working engine is a problem for a (bigger) sailing ship. Sailing into an (unknown) (sea) port is usually hard and on the open sea your engine is an important safety measure in case of problems.
  • :-) Fortunately, Atlantis, one of our companion ships was only a few miles ahead (after having crossed most the North Sea through different routes, many miles apart). We sailed towards the port of New Haven, where they towed us into the port.
  • :-( We diagnosed that there was a problem in the fuel supply, but couldn't find the real cause. We spent the sunday in the thrilling harbour town of New Haven, waiting for a mechanic to arrive.
  • :-) On monday morning, a friendly mechanic called Neil fixed our fuel tank - it turned out there was a hose in there with a clogged filter as well as a kink near the end that prevented fuel from getting through. Neil cut off the end of the tube, fixing both problems at once and our engine ran again like a charm.

Fort in the Solent

  • :-( We continued towards Wight around noon, but shortly after passing No mans land Fort - guess what - our engine stopped running again.
  • :-( This time we had no friends around to tow us in, but there was a Dutch ship (the Abel Tasman) nearby. We tried to get their attention using radio, a horn and a hand-held flare, but they misunderstood our intentions and carried on. We kept on sailing, but contacted the very much polite and helpful Solent Coast Guard by radio, to let them know we were coming in by sail and might needed a tow into the harbour.
  • :-) After we finished our conversation with the Solent Coast Guard, the Abel Tasman hailed us on the radio and offered to tow us all the way to the harbor. With the wind dropping, we gladly accepted.
  • :-) On monday evening We arrived on the Isle of Wight!
  • :-) In the harbour, our engine started again (with some initial coughing). Yay for having a working engine, but without having fixed a problem, it could break down any moment. :-( We had some theories about what the problem could be, but no real confidence in any of them.
  • :-) In the morning, we called for Tristan, a mechanic from the town Cowes where we were staying. He had a look at our engine, but since it was running again, there wasn't much he could do. He found a connection in the fuel line that wasn't very tight - guessing that that had allowed some air to slip into the fuel line. He tighted the connection and left us again.
  • :-( Shortly after he left, we started the engine for a short refueling trip - and it stopped running again. We called for the mechanic to have another look, but he didn't show up all day...

On the open sea

  • :-( Near the end of the day, we didn't want to wait anymore and tried to bleed out the air from our engine, thinking that was the last problem. No matter how hard we tried, air nor fuel would come through. Debugging further, we found that the fuel tank would again not supply any fuel at all. We "fixed" this by disconnecting the fuel line, blowing into it and then sucking at it (credits to Jan-Peter for sucking up a mouthful of Diesel...). Together with the diesel there came also a wad of something that looked like mud.
  • :-( Previously, we didn't suspect any pollution of our fuel tank, because the looking glass in the fuel filter was squeeky clean. Now, prompted by the gunk coming out of the tank, we opened up the filter to have a more thorough look and it turned out the filter was nearly completely filled with the same muddy stuff (which turned out to be "Diesel Bug" - various (byproducts of) bacteria and other organisms that like liveing in diesel tanks)
  • :-) We didn't have a working engine yet, but at least we knew with confidence what the cause of our problems was, and had been. As Danny stated it: At least it is truely broken now. All we needed now was machinery to filter our fuel and rinse the inside of the tank.
  • :-( On Tuesday, we found another mechanic who couldn't help us directly, but made a few phone calls. He found someone with the exact machinery we needed. The only problem was that the machinery was in the Swanwick Marina, on the UK mainland.

Improvised fuel tank

  • :-) On Wednesday, we tried to get into contact with the mechanic in Swanwick, to see if he could come over to Wight by ferry. We weren't hopeful, so in the meantime we also found a small plastic fuel tank (normally used for larger outboard motors) that could serve as a spare diesel tank. After also finding a few meters of diesel hose, getting an extra hose barb fitted and even getting a valve for fast on-the-fly switching between tanks (Thanks to Power Plus Marine Engineering for going out of their way to help us!), we had a working engine again by the end of the day!

    We weren't confident our shiny new tank to sail all the way back to the Netherlands using it, but it would be more than sufficient to get over to Swanwick to get our main tank cleaned. As an extra bonus, it would be a good backup tank in case we still had problems after cleaning.

  • :-) Early Thursday morning, we headed over to Swanwick where Geoffrey and Johnny laboured all morning to get a ton of sludge out of our tank. They sprayed the fuel back in using a narrow piece of pipe to clean the stuff off the edges of our tank as good as possible.
  • :-( As soon as they were done, we cast off and set course to Brighton Eastbourne Dover Ramsgate Dunkerque, Oostende, Zeebrugge, Dunkerque. Unfortunately, the wind direction was unfavorable (and later also didn't abide by the prediction, in our disadvantage of course), so we ended up sailing for only a small bit and using our engine most of the way.

    We changed our destination a few times along the way, usually because we thought "We might as well go on for a bit more, the further we go now, the easier it will be for the last stretch". Near Dunkerque, we planned to go on to Zeebrugge, but the wind picked up significantly. With the waves and wind against, we were getting a very rough ride, so we stopped at Dunkerque after all.

  • :-( We came into port friday after office hours, so there was no personell anymore. In the restaurant we managed to get an access card so we could get to the toilet and back again, but the card had no balance, so we couldn't even take a shower...
  • :-( Early saturday morning we carried on, this time with favorable wind direction, but not enough wind to make a decent speed. Since we lost all the slack in our schedule already (we had planned to be home by saturday afternoon), we used our engine again to make some speed.
  • :-( Shortly after crossing the Westerschelde, our engine stopped running again (can you believe that?). We quickly found that the engine itself would run fine, but would stop as soon as you put it into gear - there was something blocking up our screw propeller. Using Danny's GoPro underwater camera we confirmed that there was indeed something stuck in the screw. We didn't want to go into the water to fix it on open sea because of the safety risk, as well as the slim chance that we'd actually be able to cut out whatever was stuck.

  • :-) We contacted the coast guard again (the dutch one this time), telling them we'd sail towards the Roompot locks and would probably need a tow through the locks and into the harbour. We contacted the Burghsluis harbour and found they had a boat crane and were willing to operate it during the evening, so we had our target.

  • :-( Then the wind dropped to nearly zero and we just floated around the North Sea for a bit. After an hour of floating, we contacted the coast guard and asked them to tow us in all the way - we were losing time quickly and with unreliable wind, there was no way we could safely cross the northern exit channel of the Westerschelde.

Diver on the open sea

  • :-) While floating around a bit more, waiting for the coast guard to arrive, we slowly floated nearer to a small boat that we had previously seen in the distance already. Now, and this is where the story becomes even more unlikely, this boat was running a "diver flag", meaning there was a diver in the water nearby.

    We slowly approached the boat and shortly after yelling over our problem to the woman on the boat, the diver emerged from the water. She told him of our problem, he swam over and under our boat and not even 30 seconds later he emerged again carrying a fairly short but thick piece of rope that had been stuck in our screw.

    We started our engine and were on our way again, suffering about a 3-4 hour delay in total.

  • :-) We contacted the coast guard again to call them off, but they were already nearly there. To resolve some formalities, they continued heading our way. A minute or so later, the Koopmansdansk, a super-speedy boat from the KNRM came alongside. After we gave (read: Tried to yell over the roaring of their motor) them some contact information and the gave us a flyer about sponsoring the KNRM, we both went our separate ways again.


  • :-( During the night, a thunderstorm came over us. I missed most of it because I was sleeping, but Jan Peter and Danny were completely drenched and attested that it was both heavy and somewhat scary due to the lightning.
  • :-) Early sunday morning we went through the locks in IJmuiden, passed through Amsterdam and the Markermeer pretty uneventful. It even seemed like I would be at home just in time to head over to Enschede for our regular Glee night with some friends.
  • :-( We were running low on diesel, so we had to find a place to fuel up in Lelystad. We found a harbour that was documented to have a gas pump, but didn't have one, so wasted some time finding one...
  • :-( When we got to the locks of Roggebotsluis, it took a long time for the bridge to open. After floating around for 10 minutes or so, we noticed that one of the boom barriers (slagboom) looked warped and bent. Turns out just before we arrived there, some car got locked in between the barriers, panicked, hit the barrier while backing up and then sped away past the barrier on the opposite side. A mechanic was being summoned, but being a sunday...
  • :-) After a short while, the bridge controller fellow came out, pulled the boom somewhat back into shape a bit and announced they'd give it a go. Thankfully for us, everything still worked and we were on our way!
  • :-) No more delays after that. Brenda picked me up in the harbour, where I took a very quick shower before we headed to Enschede. Just in time for Glee, though I started falling asleep before we were even halfway. Apparently I was exhausted...

Overall, this was a crazy journey. All the problems we had have been a bit stressful and scary at times, but it was nice nonetheless. I especially enjoyed trying get things working again together, it was really a team effort.

Improvised Arduino NMEA multiplexer

Also interesting, especially for me, is that we had installed an AIS receiver on-board, which allows receiving information from nearby ships about their name, position, course and speed. This is especially useful in busy areas with lots of big cargo ships and to safely make your way through the big shipping lanes at sea.

This AIS receiver needed to be wired up to the existing instruments and to my laptop, so we could view the information on our maps. So I spent most of the first day belowdecks figuring this stuff out. I continued fiddling with these connections and settings to improve the setup, running into some bugs and limitations. Knowing I might have to improvise, I had packed a few Arduinos, some basic electronics and an RS232 transceiver, which allowed me to essentially build a NMEA multiplexer that can forward some data and drop other data. Yay for building your own hardware and software while at sea. Also yay for not getting seasick easily :-p

Now, I'm getting ready for our next trip. We start, and possibly end, in Vlissingen this time (to buy a bit more time) and intend to go to Cherbourg. Let's hope this journey is a bit less "interesting" than the previous one!

0 comments -:- permalink -:- 19:37
Showing 1 - 10 of 161 posts
Copyright by Matthijs Kooijman