Install Raspberry Pi’s vcgencmd on Fedora

Updated: 5 minutes to read

This post is a continuation of my previous one about setting up a cluster of Raspberry Pis running Fedora. After I got the cluster to compute something, @ColsonXu, the cluster’s owner, asked me if I could monitor the CPU temperature of each Raspberry Pi by running this command:

$ /opt/vc/bin/vcgencmd measure_temp

The vcgencmd program is included in Raspberry Pi OS (formerly called Raspbian) as a utility for retrieving information about Raspberry Pi’s hardware. However, it is not included in Fedora’s software repositories. Luckily though, the source code of vcgencmd, along with the entire userland package that contains the program, is available, so we can compile it on our own.

Build and Install the Program

  1. Install the required compilers and build tools, and Git for retrieving the source code.

    $ sudo dnf install cmake gcc gcc-c++ make git

  2. Clone the userland package’s source code, then enter its directory.

    $ git clone https://github.com/raspberrypi/userland.git
$ cd userland

  3. Use ./buildme --aarch64 to compile the program for the aarch64 architecture Fedora runs on and install it.

    After the compilation completes and before the installation, there might a prompt from sudo asking for authentication. Enter the required credentials to proceed.

    $ ./buildme --aarch64

After this command completes, the vcgencmd program can be found under /opt/vc/bin.

Tell the System About /opt/vc

If the vcgencmd program is invoked now, the following error message is expected to show up:

$ /opt/vc/bin/vcgencmd
/opt/vc/bin/vcgencmd: error while loading shared libraries: libvchiq_arm.so: cannot open shared object file: No such file or directory

This error is caused by the operating system not knowing where to find the shared object file libvchiq_arm.so. That file does exist under /opt/vc/lib, but the system is not told to find the file from that directory. To solve this issue, create a file whose name ends with .conf under the /etc/ld.so.conf.d directory, and add the following line to the file:

/opt/vc/lib

Then, run the following command to apply the change:

$ sudo ldconfig

Running /opt/vc/bin/vcgencmd now should no longer produce the error message.

It can be tedious when the full path to the vcgencmd program must be entered to run it every time. To avoid this, add /opt/vc/bin to the PATH environment variable. One way of doing this is to edit ~/.bashrc:

  # User specific environment
  if ! [[ "$PATH" =~ "$HOME/.local/bin:$HOME/bin:" ]]
  then
      PATH="$HOME/.local/bin:$HOME/bin:$PATH"
  fi
+ PATH="/opt/vc/bin:$PATH"
  export PATH

Then, run the following command for the change to take effect:

$ source ~/.bashrc

Configure Device Permissions and User Group

At this point, if vcgencmd is run using a normal user account to read hardware information, the VCHI initialization failed error might show up.

Most solutions to this issue posted online would say that the user account used to run the program should be added to the video group. However, these solutions assume that vcgencmd is being run on Raspberry Pi OS. On Fedora, doing only this will not suffice. The VCHI device also needs to be configured so that video group users can access it. This is done by adding a new udev rule:

KERNEL=="vchiq",GROUP="video",MODE="0660"

This rule has only been tested on relatively-new Linux kernel versions (5.16 and above); it might not work on older kernels. If the rule does not work because the kernel version is too old, please upgrade to the latest kernel.

To add the udev rule, create a new file whose file extension is .rules under /etc/udev/rules.d, and add the rule as a line of text to the file. This can be achieved by running the following command, which will install the rule to a file called 92-local-vchiq-permissions.rules:

$ sudo tee /etc/udev/rules.d/92-local-vchiq-permissions.rules <<< 'KERNEL=="vchiq",GROUP="video",MODE="0660"'

Once the udev rule is copied to the correct location, it may be applied immediately without a reboot by using udevadm:

$ sudo udevadm trigger /dev/vchiq

To see if the rule is in effect, check the permission settings for the VCHI device file /dev/vchiq. If its group is video, then the udev rule has been successfully activated.

$ ls -l /dev/vchiq
crw-rw----. 1 root video 511, 0 Nov  9 23:17 /dev/vchiq

Once this is done, any user in the video group can invoke vcgencmd without getting the same error. The following command can be used to add the current user account to the video group; however, the change will not take effect until the account is logged out.

$ sudo usermod -aG video $USER

Use DNF to Install the Program

Don’t want to build vcgencmd by yourself? I have made an RPM package for userland and uploaded it to a Copr repository, so anyone can get vcgencmd working by simply running the following DNF commands:

$ sudo dnf copr enable leo3418/raspberrypi-userland
$ sudo dnf install raspberrypi-userland

With this installation method, all the building and installation steps described above can be skipped, including creating a .conf file under /etc/ld.so.conf.d, modifying ~/.bashrc, and installing the udev rule. The only thing that must be done manually is to add the user account to the video group.

You can also build the RPM packages for userland by yourself from the SPEC file I wrote for it. For this one, I will only give a demo of how to build the RPM packages instead of detailed instructions.