Category Archives: Hacks


I saw the recent announcements from Mozilla, Cloudflare and Google about running a trial to try and make DNS name resolution more secure.

The basic problem is that most users get their DNS server set via DHCP which is controlled by who ever runs the network (at home this tends to be their ISP, but when using public wifi this could be anybody). The first approach to help with this was Google’s public DNS service (followed by the IBM’s and Cloudflares This helps if people are technically literate enough know how to change their OS’s DNS settings and fix them to one of these providers. Also DNS is UDP based protocol which makes it particularly easy for a bad actor on the network to spoof responses.

The approach the 3 companies are taking is to run DNS over an existing secure protocol, in this case HTTPS. From Firefox version 60 (currently in beta) it is possible to set it up to do name host name resolution via DNS-Over-HTTPS.

There are currently 2 competing specifications for how to actually implement DNS-Over-HTTPS.

DNS Wireformat

This uses exactly the same data structure as existing DNS. Requests can be made via a HTTP GET or POST. For a POST the body is the binary request and the Content-Type is set to application/dns-udpwireformat.

For GET requests the payload is BASE64 encoded and passed as the dns query parameter.

In both cases the response is the same binary payload as would be made by a normal DNS server.

This approach is currently covered by this draft RFC


For this approach the request are made as a HTTP GET request with the hostname (or IP address) being passed as the name and the query type being passed as the type query parameters.

A response looks like this:

    "Status": 0,
    "RA": true,
    "RD": true,
    "TC": false,
    "AD": false,
    "CD": true,
    "Additional": [],
    "Answer": [
            "TTL": 86400,
            "data": "",
            "name": "",
            "type": 1
    "Question": [
            "name": "",
            "type": 1

With a Content-Type of application/dns-json

You can find the spec for this scheme from Google here and Cloudflare here.

Both of these schemes have been implemented by both Google and Cloudflare and either can be used with Firefox 60+.

Privacy Fears

There has already been a bit of a backlash against this idea, mainly around privacy fears. The idea of Google/CloudFlare being able to collect information about all the hosts your browser resolves scared some people. Mozilla has an agreement in place with CloudFlare about data retention for the initial trial.

Given these fears I wondered if people might still want to play with DNS-Over-HTTPS but not want to share data with Google/Cloudflare. With this in mind I thought I’d try and see how easy it would be to implement a DNS-Over-HTTPS server. Also people may want to try this out on closed networks (for things like performance testing or security testing).

It turned out not to be too difficult, I started with a simple ExpressJS based HTTP server and then started to add DNS support. Initially I tried a couple of different DNS NodeJS nodes to get all the require details and in the end settled on dns-packet and actually sending my own UDP packets to the DNS server.

I’ve put my code up on github here if anybody wants a play. The should include details about how to set up Firefox to use an instance.

Logging request & response body and headers with nginx

I’ve been working a problem to do with oAuth token refresh with the Amazon Alexa team recently and one of the things they have asked for is a log of the entire token exchange stage.

Normally I’d do this with something like Wireshark but as the server is running on a Amazon EC2 instance I didn’t have easy access to somewhere to tap the network so I decided to look for another way.

The actual oAuth code is all in NodeJS + Express but the whole thing is fronted by nginx. You can get nginx to log the incoming request body relatively simply, there is a $request_body variable that can be included in the logs, but there is no equivalent $resp_body.

To solve this I turned to Google and it turned up this answer on Server Fault which introduced me to the embedded lua engine in nginx. I’ve been playing with lua for some things at work recently so I’ve managed to get my head round the basics.

The important bit of the answer is:

lua_need_request_body on;

set $resp_body "";
body_filter_by_lua '
  local resp_body = string.sub(ngx.arg[1], 1, 1000)
  ngx.ctx.buffered = (ngx.ctx.buffered or "") .. resp_body
  if ngx.arg[2] then
     ngx.var.resp_body = ngx.ctx.buffered

I also wanted the request and response headers logging so a little bit more lua got me those as well:

set $req_header "";
  set $resp_header "";
  header_filter_by_lua ' 
  local h = ngx.req.get_headers()
  for k, v in pairs(h) do
      ngx.var.req_header = ngx.var.req_header .. k.."="..v.." "
  local rh = ngx.resp.get_headers()
  for k, v in pairs(rh) do
      ngx.var.resp_header = ngx.var.resp_header .. k.."="..v.." "

This combined with a custom log format string gets me everything I need.

log_format log_req_resp '$remote_addr - $remote_user [$time_local] '
'"$request" $status $body_bytes_sent '
'"$http_referer" "$http_user_agent" $request_time req_header:"$req_header" req_body:"$request_body" resp_header:"$resp_header" resp_body:"$resp_body"';

Native NodeJS callbacks with Context

As I mentioned back in September I’ve recently started a new job. Due to the nature of the work I’m not going to be able to talk about much of it. But when there are things I can I’ll try and write them up here.

One of my first tasks has been to write a Node-RED wrapper around a 3rd party native library. This library provides a 2 way communication channel to a prototyping environment so needs to use threads to keep track of things in both directions and make use of callbacks to pass that information back into the Javascript side. I dug around for some concrete examples of what I was trying and while I found a few things that came close I didn’t find exactly what I was looking for so here is a stripped back version of the node I created to use as a reference for later.

This is the C++ method that is called when a new instance of the native node module is created. It takes an object reference as an argument to be stored away and used as the context for the callback later.

void Test::New(const Nan::FunctionCallbackInfo<v8::Value>& info) {
  if (info.IsConstructCall()) {
    // Invoked as constructor: `new MyObject(...)`
    Test* obj = new Test();

    v8::Local<v8::Object> context = v8::Local<v8::Object>::Cast(info[0]);

    uv_loop_t* loop = uv_default_loop();
    uv_async_init(loop, &obj->async, asyncmsg);
  } else {
    // Invoked as plain function `MyObject(...)`, turn into construct call.
    const int argc = 2;
    v8::Local<v8::Value> argv[argc] = { info[0], info[1] };
    v8::Local<v8::Function> cons = Nan::New<v8::Function>(constructor);

The object is created like this on the javascript side where the this is the reference to the object to be used as the context:

function Native() {;
  //passes "this" to the C++ side for callback
  this._native = new native.Test(this);

We then make the callback from C++ here:

void Test::asyncmsg(uv_async_t* handle) {
  Nan::HandleScope scope;

  //retrieve the context object
  Test* obj = (Test*)((callbackData*)handle->data)->context;
  v8::Local<v8::Object> context = Nan::New(obj->context);

  //create object to pass back to javascript with result
  v8::Local<v8::Object> response = Nan::New<v8::Object>();
  response->Set(Nan::New<v8::String>("counter").ToLocalChecked(), Nan::New(((callbackData*)handle->data)->counter));

  v8::Local<v8::Value> argv[] = { response };

Which ends up back on the javascript side of the house here:

Native.prototype._status = function(status) {
  this.emit("loop", status.counter);

I’ve uploaded code to githib here if you want to have look at the whole stack and possibly use it as a base for your project.

Adding audio to the CCTV of your burgular

In a full on throw back to the 2nd ever post to this blog, back in February 2010, I’ve recently been updating the system that sends me a video when there is movement in my flat via MMS and email.

I thought I’d try and add audio to the video that gets sent. A quick Google turned up two options, one was to use the sox command and it’s silence option, the second uses the on_event_start triggers in motion as a way to record the audio at the same time as capturing the movement video. I went with the second option and tweaked it a bit to make it pick the right input for my system and to direct encode the audio to MP3 rather than WAV to save space.

on_event_start arecord --process-id-file /var/motion/ -D sysdefault:CARD=AF -f S16_LE -r 22050 - | lame  -m m - /var/motion/%Y%m%d_%H%M%S.mp3

The other useful addition was the –process-id-file /var/motion/ which writes the process id to a file so I can just use this to stop the recording rather than having to use grep and awk to find the process in the ps output.

on_event_end kill `cat /var/motion/`

Now it’s just a case of combining the video from motion with the audio. I can do this with ffmpeg when I re-encode the video into the 3gp container to make it suitable for sending via a MMS message.

ffmpeg -i movement.avi -i movement.mp3 -map 0:v -map 1:a -c:a aac -strict -2 -s qcif -c:v h263 -y /var/www/html/cam1/intruder.3gp

This seams to work but the output file is a little large. The default audio encoding seams to be at 160k bitrate, I wound it down to 32k and the file size got a lot better.

ffmpeg -i movement.avi -i movement.mp3 -map 0:v -map 1:a -c:a aac -b:a 32k -strict -2 -s qcif -c:v h263 -y /var/www/html/cam1/intruder.3gp

I’d like to try the AMR audio codec but I can’t get ffmpeg to encode with it at the moment, so I’m just going to email the mp3 of the audio along with the high res AVI version of the video and just send the low res 3GP version via MMS.

Fist pass TRÅDFRI MQTT Bridge

I’ve been working on integrating the new IKEA TRÅDFRI Lights into my Home Automation system. I’d really like a native NodeJS system so I can plug it directly into Node-RED, but I’ve not found a working CoAP over DTLS setup just yet.

So in the mean time I’ve been working on a very basic MQTT to CoAP client bridge in Java using the Eclipse Californium library.

It still needs some work, but here is the first pass:


import java.util.HashMap;
import java.util.logging.Level;

import org.eclipse.californium.core.CaliforniumLogger;
import org.eclipse.californium.core.CoapClient;
import org.eclipse.californium.core.CoapHandler;
import org.eclipse.californium.core.CoapResponse;
import org.eclipse.californium.core.Utils;
import org.eclipse.californium.core.coap.MediaTypeRegistry;
import org.eclipse.californium.scandium.DTLSConnector;
import org.eclipse.californium.scandium.ScandiumLogger;
import org.eclipse.californium.scandium.config.DtlsConnectorConfig;
import org.eclipse.californium.scandium.dtls.pskstore.StaticPskStore;
import org.eclipse.paho.client.mqttv3.IMqttDeliveryToken;
import org.eclipse.paho.client.mqttv3.MqttCallback;
import org.eclipse.paho.client.mqttv3.MqttClient;
import org.eclipse.paho.client.mqttv3.MqttException;
import org.eclipse.paho.client.mqttv3.MqttMessage;
import org.eclipse.paho.client.mqttv3.persist.MemoryPersistence;
import org.json.JSONArray;
import org.json.JSONObject;

 * @author hardillb
public class Main {
  static {
//    ScandiumLogger.initialize();
//    ScandiumLogger.setLevel(Level.FINE);
  private DTLSConnector dtlsConnector;
  private MqttClient mqttClient;
  private CoapEndpoint endPoint;
  private String ip;
  private HashMap<String, Integer> name2id = new HashMap<>();
  Main(String psk, String ip) {
    this.ip = ip;
    DtlsConnectorConfig.Builder builder = new DtlsConnectorConfig.Builder(new InetSocketAddress(0));
    builder.setPskStore(new StaticPskStore("", psk.getBytes()));
    dtlsConnector = new DTLSConnector(;
    endPoint = new CoapEndpoint(dtlsConnector, NetworkConfig.getStandard());
    MemoryPersistence persistence = new MemoryPersistence();
    try {
      mqttClient = new MqttClient("tcp://localhost", MqttClient.generateClientId(), persistence);
      mqttClient.setCallback(new MqttCallback() {
        public void messageArrived(String topic, MqttMessage message) throws Exception {
          // TODO Auto-generated method stub
          System.out.println(topic + " " + message.toString());
          String parts[] = topic.split("/");
          int id = name2id.get(parts[1]);
          String command = parts[3];
            JSONObject json = new JSONObject("{\"9001\":\"Living Room Light\",\"9020\":1491515804,\"9002\":1491158817,\"9003\":65537,\"9054\":0,\"5750\":2,\"3\":{\"0\":\"IKEA of Sweden\",\"1\":\"TRADFRI bulb E27 opal 1000lm\",\"2\":\"\",\"3\":\"\",\"6\":1},\"9019\":1,\"3311\":[{\"5850\":1,\"5851\":10,\"9003\":0}]}");
            JSONObject settings = json.getJSONArray("3311").getJSONObject(0);
            if (command.equals("dim")) {
              settings.put("5851", Integer.parseInt(message.toString()));
            } else if (command.equals("on")) {
              if (message.toString().equals("0")) {
                settings.put("5850", 0);
                settings.put("5851", 0);
              } else {
                settings.put("5850", 1);
                settings.put("5851", 128);
            String payload = json.toString();
            Main.this.set("coaps://" + ip + "//15001/" + id, payload);
          } catch (Exception e) {
        public void deliveryComplete(IMqttDeliveryToken token) {
          // TODO Auto-generated method stub
        public void connectionLost(Throwable cause) {
          // TODO Auto-generated method stub
    } catch (MqttException e) {
      // TODO Auto-generated catch block
  private void discover() {
    try {
      URI uri = new URI("coaps://" + ip + "//15001");
      CoapClient client = new CoapClient(uri);
      CoapResponse response = client.get();
      JSONArray array = new JSONArray(response.getResponseText());
      for (int i=0; i<array.length(); i++) {
        String devUri = "coaps://"+ ip + "//15001/" + array.getInt(i);;
    } catch (URISyntaxException e) {
      // TODO Auto-generated catch block
  private void set(String uriString, String payload) {
    System.out.println("payload\n" + payload);
    try {
      URI uri = new URI(uriString);
      CoapClient client = new CoapClient(uri);
      CoapResponse response = client.put(payload, MediaTypeRegistry.TEXT_PLAIN);
      if (response.isSuccess()) {
      } else {
    } catch (URISyntaxException e) {
      // TODO Auto-generated catch block
  private void watch(String uriString) {
    try {
      URI uri = new URI(uriString);
      CoapClient client = new CoapClient(uri);
      CoapHandler handler = new CoapHandler() {
        public void onLoad(CoapResponse response) {
          JSONObject json = new JSONObject(response.getResponseText());
          if (json.has("3311")){
            MqttMessage message = new MqttMessage();
            int state = json.getJSONArray("3311").getJSONObject(0).getInt("5850");
            String topic = "TRÅDFRI/" + json.getString("9001") + "/state/on";
            String topic2 = "TRÅDFRI/" + json.getString("9001") + "/state/dim";
            name2id.put(json.getString("9001"), json.getInt("9003"));
            MqttMessage message2 = new MqttMessage();
            int dim = json.getJSONArray("3311").getJSONObject(0).getInt("5851");
            try {
              mqttClient.publish(topic, message);
              mqttClient.publish(topic2, message2);
            } catch (MqttException e) {
              // TODO Auto-generated catch block
          } else {
            System.out.println("not bulb");
        public void onError() {
          // TODO Auto-generated method stub
    } catch (URISyntaxException e) {
      // TODO Auto-generated catch block

   * @param args
  public static void main(String[] args) throws InterruptedException {
    String psk = args[0];
    String ip = args[1];
    Main m = new Main(psk, ip);;


I’ve tagged the code onto the gist as well for now, but I’ll check the whole thing in as a separate project soon.

EDIT: now with it’s own Github repo here

More improvements to the WeMo Node-RED Nodes

Having got all my bulbs working again with my WeMo Link device it was time to have another play with the Node-RED WeMo nodes.

  • First up close the gap between what the doc claims the nodes do and what they actually do. This is mainly the WeMo Out node which claimed to support an input object that could set both the state and the brightness at the same time for light bulbs and groups. In actuality it only set the state. This update fixes this along with supporting setting colour and colour temperature as well if the bulb supports those capabilities.
  • The WeMo in node (event node) now includes the capability name as well as it’s code when a bulb or group changes.
  • Fixed Light groups to actually work
  • Automatically set the node name to the discovered device name to stop you having to set it via the name field in the config

The biggest change is the addition on the new WeMo Lookup node. This node queries a given device for it’s current state.

For Sockets the node sets the msg.payload to something very similar to the event node’s staus field, so 0 for off and 1 for on (and 8 for on/standby in the case of the insight socket.). For lights/light groups it outputs an object similar to the Event node, with keys for each capability the light/group has. At the moment the color field is still in X,Y values not RGB.

This makes it possible to implement flows that carry out relative changes without having to keep a permanent record of the state of the device in the context. This let’s you do fun things like this:

Powermate dimmer with Node-RED

This flow looks up the current brightness level and then increases or decreases it based on the direction the Powermate is turned and toggles the light on/off when it’s pressed.

Version 0.1.11 should have gone live on today.

TRÅDFRI – Ikea’s new Smart lighting system

This week Ikea announced a new set of Smart Lighting products called TRÅDFRI (translates to ‘Wire Free’).

Reading the various articles it sounds like it’s a Zigbee Light Link based system so I was interested in having a play to see if I could make the bulbs work with the WeMo kit I already have, but also given how cheap the gateway is I thought I’d grab one of those as well to see if I could work out what the network protocol is so I can write a similar Node-RED node to the existing WeMo one.

A (remarkably) short trip to the local branch and I came a way with 2 items:

A E27 Blub that can be dimmed and a remote dimmer £15

E27 LED Bulb + Dimmer
E27 LED Bulb + Dimmer

The dimmer looks very cool. It looks like it’s accelerometer based, comes with a little magnetic mount that can be stuck or screwed to the wall or it can be just stuck to the front of the fridge. If you turn it slowly it dims/brightens or if you turn it faster it turns the light on and off. You can pair each remote with up to 10 bulbs, so it can control a whole room (all be it all at once).

Bulbs on their own start at £9 and they are doing E14, GU10 formats as well as several versions of E27.

A Network Gateway that plugs into the router £25

TRÅDFRI Ethernet bridge
TRÅDFRI Ethernet bridge

The bridge is USB powered and comes with little USB power supply and a length of ethernet cable to connect it to your router.

I’m going to set the bridge up on a separate network and capture the traffic between the Android App and the bridge to see what the network traffic looks like. I’ll post again with the captured data and my progress as well as sticking it all up on github so others can build libraries for other systems, like OpenHab.

I’m keeping the notes on how I’m getting on with working out how to drive the gateway here


It’s finally ready. I’ve been working on a Node-RED node to act on Amazon Alexa Home Skill directives since November last year. The skill was approved some time very early this morning and now should be available in the UK, US and Germany.

I’ll be mailing all the folks that have already signed up some time later today to let them know they can finally start using the skill, but for the rest of you here is a brief introduction (full details in earlier post).

Alexa Home Skill’s allow you to say the much more natural “Alexa, turn on the kitchen light” rather than “Alexa, ask Jeeves to turn on the kitchen light”, where “Jeeves” is the name of skill you have to remember. Some of the basic commands are:

  • Turn On/Off
  • Dim/Brighten
  • Set/Get Temperature
  • Lock/Unlock

With this node and service you can wire those commands to nearly anything you can control via Node-RED.

Node-RED - Alexa Smart Home Skill

You can install the node with the following commands:

cd ~/.node-red
npm install node-red-contrib-alexa-home-skill

Or via the Manage Palette option in the Node-RED editor.

If you have already installed this module please make sure you update to the latest version (0.1.13) to get the best support for all the voice commands.

There are detailed instructions on how to set everything up here.

Here is an example flow using the node. This turns a light on then automatically turns it off after 5mins. It uses the switch node to detect if it’s a request to turn the light on or off. When following the On branch it uses a trigger node to first send a payload of true then, 5 minutes later it sends false to the WeMo node.

On then Auto Off flow

This sort of flow would be great for a set of outside lights or maybe an electric heater. I also have some updates to the node-red-nodes-wemo package to make dimming/brightening by specific amounts easier, I’ll try and get them out by the weekend.

If you have problems with this node please do not post comments here, it really isn’t the best place to work issues. Open a issue on github here then it can be properly tracked.

Making the WeMo Link work with 3rd party bulbs

innr bulb
I’ve been playing with my WeMo Link device again. A couple of weekends ago I made a mistake, I used the Belkin WeMo Android app to turn a light on (rather than ask Alexa to turn it on via Node-RED). While doing this the app suggested I update the firmware on the Link device. I decided to let it do the update, this led to whole host of issues with the device not wanting to connect to the WiFi.

In the end I had to reset the device, and re-pair the bulbs. The re-pairing worked, but the Osram bulbs didn’t show up in the WeMo android app any more. It turns out the app no longer show bulbs which haven’t been “WeMo Certified”. They still show up when you query the API directly and I can control them via my Node-RED nodes but I couldn’t add the bulbs to groups. The WeMo Link supports the ZigBee Light Link protocol just like Osram Lightify, Philips Hue and the Innr lights so there is no reason why all of these things shouldn’t be able to play nicely with each other.

This sent me back to reverse engineering the SOAP API to interact with the WeMo Link.

I’ve had a pretty good go at working out the protocol already, resulting in the wemo-control.js script and the Node-RED WeMo nodes, but this is just basic discovery and control not really “admin” tasks. I decided to break this work out into it’s own script.

There are 2 main tasks that this script will have to do, add new bulbs and create groups.

Adding bulbs

This was actually pretty easy, the API end points were pretty obvious in the list.

  • OpenNetwork – allows bulbs to join the mesh
  • GetEndDevices – when used with the UNPAIRED_LIST filter it shows just the new bulbs.
  • IdentifyDevice – makes a bulb flash so you can work out which bulb is which if you discover new bulbs
  • AddDevice – adds a new bulb to the mesh
  • CloseNetwork – stops bulbs joining the mesh

None of these calls take any complex arguments and all are available either via discovery responses or other simple calls. If you chain them together in the order above you end up with your new bulbs available (at least to my scripts even if not in the WeMo app).

Creating Groups

This one was a little harder, while the API list a CreateGroup endpoint, it says that it takes as a single argument if type ReqCreateGroup which is listed to be a string. Now from experience I can guess that this string is actually a URL encoded XML fragment. There are no hints as to what this XML fragment might look like. This led to a slight diversion down a rabbit hole to set up a raspberry pi as a WiFi AccessPoint bridged on to my local network so I could run tcpdump to make sure I captured all sides of the conversation between my tablet and the WeMo Link. A little bit of formatting and collating in Wireshark and we hit pay dirt:

<?xml version="1.0" encoding="UTF-8"?>
    <GroupName>Lighting Group</GroupName>

The full capture can be found here.

It all looks pretty self explanatory,

  • A unique id for the group (looks like it’s epoch time in seconds)
  • Name of the group (probably should be XML escaped, but we’ll keep them simple)
  • The list of device IDs to include in the group
  • The subset of capabilities that all the devices in the group support
  • Some starting values for those capabilities

Now I had the format of the messages I need to send it’s time to actually write some code. The first pass is up here, it’s a little rough and ready but I’ll try and clean it up a bit later and add a command to rename devices, but it’s Friday afternoon and I’m typing this up in a bar…

To add a bulb you would follow this flow of commands:

node bridge.js open
node bridge.js unpaired
<?xml version="1.0" encoding="utf-8"?><DeviceLists><DeviceList><DeviceListType>Unpaired</DeviceListType><DeviceInfos><DeviceInfo><DeviceIndex>5</DeviceIndex><DeviceID>00158D0001696252</DeviceID><FriendlyName>OnOff Light</FriendlyName><IconVersion>1</IconVersion><FirmwareVersion>01</FirmwareVersion><CapabilityIDs>10006,10008,30008,30009,3000A</CapabilityIDs><CurrentState>,,,,</CurrentState><Manufacturer>innr</Manufacturer><ModelCode>BY 165</ModelCode><productName>lighting</productName><WeMoCertified>NO</WeMoCertified></DeviceInfo></DeviceInfos></DeviceList></DeviceLists>
node bridge.js indentify <device id>
node bridge.js add <device id>
node bridge.js close

To create a new group:

node bridge.js createGroup "Group Name" "00158D0001696252,00158D0001696253" "10006,10008,30008,30009,3000A", "0,255:0,,,"

Updated Pi Zero Gadgets

Following on from my last post I’ve continued to work on improving my instructions for a USB connectable gadget based on a Raspberry Pi Zero.

Firstly I’ve got a slight improvement to the dnsmasq config I mentioned last time. This removes the dnsmasq.leases file which can cause issues if you plug the Zero into multiple computers. This can be a problem because while I had managed to fix the mac address for Host computer end of the connection the OS still pushes down the host name and it’s own unique id when it requests a DHCP address from dnsmasq, this causes dnsmasq to cycle through it’s small pool of addresses. This combined with the fact the clock on Zero is not battery backed up so only gets set correctly when it can reach internet can cause strangeness with addresses getting expired in strange ways. Anyway there is a pretty simple fix.

Adding leasefile-ro to the dnsmasq config causes it to not write lease information to disk, but rely on the dhcp-script to keep track of things. To do this I needed to add handling for a new option to the script to deal with dnsmasq wanting to read the current lease state at startup.


if [[ $op == "init" ]]; then
  exit 0

if [[ $op == "add" ]] || [[ $op == "old" ]]; then
  route add default gw $ip usb0

Now on to getting things working better with Windows machines.

To do this properly we need to move from the g_ether to the g_multi kernel module, this lets the Zero be a USB Mass Storage device, a network device (and a serial device) at the same time. This is useful because it lets me bundle .inf files that tell Windows which drivers to use on the device it’s self so it they can be installed just by plugging it in.

The first order of business is to fix the cmdline.txt to load the right module, after making the changes in the last post it looks like this:

dwc_otg.lpm_enable=0 console=serial0,115200 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait modules-load=dwc2,g_ether

The g_ether needs replacing with g_multi so it looks like this:

dwc_otg.lpm_enable=0 console=serial0,115200 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait modules-load=dwc2,g_multi

next we need to fix the options passed to the module, these are in the /etc/modprobe.d directory probably in a file called g_ether.conf. We should don’t need to, but to make it obvious when we come back to this a lot later on we’ll rename it to g_multi.conf. Now we’ve renamed it we need to add a few more options.

It currently looks like this:

options g_ether host_addr=00:dc:c8:f7:75:15 dev_addr=00:dd:dc:eb:6d:a1

It needs the g_ether needs changing to e_multi and some options adding to point to a disk image.

options g_multi file=/opt/disk.iso cdrom=y ro=y host_addr=00:dc:c8:f7:75:15 dev_addr=00:dd:dc:eb:6d:a1

Now we have the config all sorted we just need to build the disk image. First we need to create a directory called disk-image to use as the root of the file system we will share from the Zero. Then we need to get hold of the 2 .inf files, they ship with the Linux Kernel doc, but can be found online (serial port, RNDIS Ethernet).

Once we have the files place them in a directory called disk-image/drivers. We should also create a README.txt to explain what’s going on, put that in the root of disk-image. Along side that we want to create a file called Autorun.inf, this tell Windows about what’s on the “cd” when it’s inserted and where it should search for the driver definitions.




Full details of what can go in the Autorun.inf file can be found here, but the interesting bits are the DriverPath=drivers which points to the directory that we put the .inf files in earlier. Also the open=”documentation\index.html” which should open documentation/index.html when the device is plugged in which explains how to install the drivers. I also added an icon file so the drive shows up looking like a clock in the file manager.

That should be the bare minimum that needs to be on the image, but I ran into an issue with the g_multi module complaining the disk image being too small, to get round this I just stuck a 4mb image in the directory as well. To build the iso image run the following command:

mkisofs -o disk.iso -r -J -V "Zero Clock" disk-image

This will output a file called disk.iso that you should copy to /opt/disk.iso on the Zero (I built the image on my laptop as it’s easier to edit files and mkisofs is not installed in the default raspbian image).

This is working well on Linux and Windows, but I’m still getting errors on OSx to do with the file system. It’s not helped by the fact I don’t have a Mac to test on so I end up having to find friends that will let me stick a random but of hardware in to the side of their MacBook.

Once I’ve got the OSx bits sorted out I’ll put together script to build images for anything you want.

So now we have a Pi Zero that should look like a CD-ROM drive and a Network adapter when plugged into pretty much any machine going, it brings the driver information with it for windows, sets up a network connection with a static IP address and a Avahi/Bonjour/mDNS address to access it. I’m planning on using this to set up my Linear Clock on the local WiFi but there are all manner of interesting things that could be done with a device like this. e.g. an offline Certificate Authority, a 2FA token, a Hardware VPN solution or a Web Controllable display device.