fwup for A/B image upgrades on raspberrypi machines with NervesCloud, part I

This blog post introduces nerves-hub-link running on raspberrypi machines in order to upgrade firmware via NervesCloud.

On previous blog posts about fwup, we have explored and tested using QEMU based machines. Now it's time to change the target and use a real machine.

Nowadays, raspberrypi is an accessible development platform for prototyping and why not for real use cases too.

On Yocto side, there is a stable layer called meta-raspberrypi which implements many of raspberrypi boards.

The end goal is to update the target board using NervesCloud, the same way that was done using QEMU based machines.

The focus of this post is about how to prepare a working Yocto setup for raspberry context. I'll not show a demonstration this time, reserving it for the next blog posts ;)

YP/OE Setup

I'll try to simplify the YP/OE setup to just tree small steps:

note

This is the same steps taken for the previous blog post called fwup for A/B image upgrades on QEMU machines with NervesCloud, part III. But now, with one additional layer called meta-raspberrypi.

Also, removing the meta-qemu-bsp.

1. Cloning all repositories for master release:

   git clone --branch master git://git.yoctoproject.org/poky
   git clone --branch master https://github.com/openembedded/meta-openembedded.git
   git clone --branch master https://github.com/fwup-home/meta-fwup
   git clone --branch master https://github.com/joaohf/meta-fwup-examples
   git clone --branch master https://github.com/meta-erlang/meta-erlang
   git clone --branch master https://github.com/meta-erlang/meta-axon
   git clone --branch master https://github.com/joaohf/meta-nerves-hub
   git clone --branch master https://github.com/agherzan/meta-raspberrypi

2. Source the init build environment script:

   cd poky
   source oe-init-build-env ../build

3. Add the needed layers:

   bitbake-layers add-layer ../meta-openembedded/meta-oe
   bitbake-layers add-layer ../meta-fwup
   bitbake-layers add-layer ../meta-fwup-examples
   bitbake-layers add-layer ../meta-nerves-hub
   bitbake-layers add-layer ../meta-raspberrypi
   bitbake-layers add-layer ../meta-axon
   bitbake-layers add-layer ../meta-erlang

Configuring the build environment

For this use case, the quickest way is edit and add the conf/local.conf configuration file.

We start enabling some multiconfig configurations:

BBMULTICONFIG ?= "\
    raspberrypi0-wifi-nerves-hub-link \
    raspberrypi4-64-nerves-hub-link \
    raspberrypi5-nerves-hub-link \
   "

The layer meta-axon provides some samples for raspberry configurations with nerves-hub enabled.

multiconfig is useful because it simplify bitbake configurations when we have many target boards. On my test setup I'm building for raspberrypi 0, 4 and 5.

Building images

Next, we need to build images:

bitbake multiconfig:raspberrypi4-64-nerves-hub-link:core-image-base

bitbake multiconfig:raspberrypi5-nerves-hub-link:core-image-base

bitbake multiconfig:raspberrypi0-wifi-nerves-hub-link:core-image-base

note

core-image-base is being used here. It's suitable for raspberrypi.

After the above commands, the result are the following files:

tmp-raspberrypi0-wifi-glibc-nerves-hub-link/deploy/images/raspberrypi0-wifi/core-image-base-raspberrypi0-wifi.rootfs.fwup
tmp-raspberrypi4-64-glibc-nerves-hub-link/deploy/images/raspberrypi4-64/core-image-base-raspberrypi4-64.rootfs.fwup
tmp-raspberrypi5-glibc-nerves-hub-link/deploy/images/raspberrypi5/core-image-base-raspberrypi5.rootfs.fwup

And are ready to flash into a bootable device.

Flashing image

Before running fwup commands:

bitbake fwup-native -caddto_recipe_sysroot

For each fwup image, and using fwup for flashing (one sdcard for each image):

rpi0:

oe-run-native fwup-native \
  fwup -a -d /dev/sdc \
  -t complete \
  -i tmp-raspberrypi0-wifi-glibc-nerves-hub-link/deploy/images/raspberrypi0-wifi/core-image-base-raspberrypi0-wifi.rootfs.fwup

rpi4:

oe-run-native fwup-native \
  fwup -a -d /dev/sdc \
  -t complete \
  -i tmp-raspberrypi4-64-glibc-nerves-hub-link/deploy/images/raspberrypi4-64/core-image-base-raspberrypi4-64.rootfs.fwup

rpi5:

oe-run-native fwup-native \
  fwup -a -d /dev/sdc \
  -t complete \
  -i tmp-raspberrypi5-glibc-nerves-hub-link/deploy/images/raspberrypi5/core-image-base-raspberrypi5.rootfs.fwup

And booting

Just put the sdcard into raspiberrypi and boot. It should work as expected.

note

I promised, for the next blog post I'll share a full demonstration

Some important details not discussed before

meta-nerves-hub

The meta-nerves-hub layer has a recipe for nerves-hub-link_2.5.2.bb. This recipe installs a standalone nerves-hub-link. I've tested the version 2.5.2 and it works for my initial goals.

However, the later nerves-hub-link versions are not prepared to run as standalone application anymore. It needs to be part of a BEAM application in order to be properly configured.

I've tested recent versions and ended up with nerves-hub-link fighting with linux network stack; for instance trying to configure it where it should not or using others ways to detect configurations.

note

It's up to you when design Yocto based linux distributions select what network manager best fits to your needs. Yocto gives to you many options as well a default one.

So, the next organic step will be to implement a BEAM application that uses nerves-hub-link. That would be a kind of nerves-hub-link-agent. I just want to start it and play nice with linux network configuration stack.

Conclusions

So, this blog post was more to make a milestone for this trip. And we are getting closer to get Yocto based linux distro full integrated with NervesCloud without thinking too much.