This guide provides steps to setup the i.MX 93 Evaluation Kit and to cross-compile an LVGL application to run it the target.

You can purchase the i.MX 93 Evaluation Kit from NXP website.

The default buffering is fbdev.

Frame buffer, 1 thread

Frame buffer, 2 threads

The other configurations that can be used are:

• DRM
• Wayland

Any of these buffering strategies can be used with multiple threads to render the frames.

• MCU:

  • Dual Cortex-A55 @1.7GHz
  • Cortex-M33 @250MHz
  • Arm® Ethos™ U-65 MicroNPU
  • EdgeLock® Secure Enclave

• RAM: 2 GB LPDDR4X / LPDDR4

  • 16 bits data bus with inline EEC
  • 3.7 GT/s

• Flash:

  • 16 GB eMMC5.1
  • MicroSD Slot

• GPU: PowerVR

• MIPI DSI (mini-SAS)
• X4 Lane LVDS (HDR)
• MIPI-CSI (Camera 22P RPi)
• Parallel RGB LCD (2x20 RPi)
• Parallel Camera (2x20 RPi)

• X2 USB C 2.0 Connectors
• 2x GbE RJ45
• CAN (HDR)
• RPi 2X20 GPIO HDR
• MFi Module Support
• X4 CH ADC Support

• JTAG Connector
• UART Via USB C
• Remote Debug Support
• Power Measurement Support

This document from NXP provides detailed information for the hardware setup. The following guide is inspired from this.

The display used in this guide is the lvds pannel, with a resolution of 1280x800.

The EMMC on the board should come flashed with an image.

• To specify the board booting from the SD card, follow the booting switch table:

  The i.MX93 used has a Cortex-A55 (use SW 4-1[0010]):

  

• The following picture shows the jumper setup for J704, J703 and J702:

• Board setup:

  

  • Connect the screen to the
  • Connect RJ45 on any ethernet port. The board must be connected on the same LAN than the host.
  • Connect USB-C power (black USB - J301)
  • (Optional) Connect USB-C debug (gray - J401)

The board should boot and the screen should display something

There are two options:

• Option 1: build Yocto image:

  • Required packages to install on host:

    sudo apt install gawk wget git diffstat unzip texinfo gcc build-essential chrpath socat cpio python3 python3-pip xz-utils debianutils iputils-ping python3-git python3-jinja2 python3-subunit zstd liblz4-tool file locales libacl1

  • (optional) If you already have the "repo utility", skip this step.

    sudo apt install repo

  • Clone the yocto project

    mkdir imx-yocto-bsp
    cd imx-yocto-bsp
    repo init -u https://github.com/nxp-imx/imx-manifest -b imx-linux-scarthgap -m imx-6.6.23-2.0.0.xml
    repo sync

  • Build the image

    # Use the script to setup the build folder and modify the conf files
    DISTRO=fsl-imx-wayland MACHINE=imx93-11x11-lpddr4x-evk source imx-setup-release.sh -b build-media
    # Build the image
    bitbake imx-image-multimedia

• Option 2: download a pre-built image: The board comes supplied with an image on the EMMC. First we replicate this setup on the SD card:

  • Download the pre-built images and binaries here. Choose the linux version and download the image for i.MX93 EVK

After downloading or building the image, flash it on the SD card:

zstdcat imx-image-multimedia-imx93-11x11-lpddr4x-evk.rootfs-20240918104911.wic.zst | sudo dd of=/dev/sda bs=1M conv=fsync status=progress

Use the correct dtb. Mount the image and on the boot partition, use the dtb called "imx93-11x11-evk-boe-wxga-lvds-panel.dtb"

# cd into the boot partition of the SD card mounted and do the following command
cp imx93-11x11-evk-boe-wxga-lvds-panel.dtb imx93-11x11-evk.dtb

This modification can also be applied using the file manager.

This guide was tested on Ubuntu 22.04 host.

• Follow this tutorial to install and setup docker on your system.

• Support to run ARM64 docker containers on the host:

  sudo apt-get install qemu-user-static
  docker run --rm --privileged multiarch/qemu-user-static --reset -p yes

sudo apt install picocom nmap

Clone the repository:

git clone --recurse-submodules https://github.com/lvgl/lv_port_nxp_imx93.git

Build the docker image and the lvgl benchmark application:

cd lv_port_nxp_imx93
./scripts/docker_setup.sh --create-image
./scripts/docker_setup.sh --build-app

Run the executable on the target:

• The board got an IP from DHCP. Both devices (the board and the host) should on the same network. Get the IP of the target board.

  • Option 1: from the UART, on the board:

    sudo picocom -b 115200 /dev/ttyUSB0
    ## Then inside the console, log as "root", no password required
    ## Then retrieve the ip of the board
    ip a

  • Option 2: Get the IP from your host with nmap.

    ## Find the IP of the board. You need to know your ip (ifconfig or ip a)
    ## HOST_IP should be built like this :
    ## If the ip is 192.168.1.86, in the following command HOST_IP = 192.168.1.0/24
    nmap -sn <HOST_IP>/24 | grep imx93-11x11-lpddr4x-evk

• Then transfer the executable on the board:

  scp lvgl_port_linux/bin/lvgl-app root@<BOARD_IP>:/root
  

• Start the application

  ssh root@<BOARD_IP>
  ######################################
  ## WARNING: do not stop these services if using wayland demo
  systemctl stop weston.socket
  systemctl stop weston.service
  ######################################
  ./lvgl-app

Some configurations are provided in the folder lvgl_conf_example . For more details about the lv_conf.h options, you can have a look to lvgl documentation

The default configuration used is lv_conf_fb_4_threads.h. To change the configuration, modify the lvgl_port_linux/lv_conf.h file with the desired configuration.

Also modify the lv_port_linux/CMakelists.txt file option:

• LV_USE_WAYLAND
• LV_USE_SDL
• LV_USE_DRM

Default is for fbdev backend. Only set 1 of these options to "ON" and ensure it's coherent with lv_conf.h. This can also be changed from the script scripts/build_app.sh.

The folder lvgl_port_linux is an example of an application using LVGL.

LVGL is integrated as a submodule in the folder. To change the version of the library:

cd lvgl_port_linux
git checkout <branch_name_or_commit_hash>

The file main.c is the default application provided and is configured to run the benchmark demo provided by LVGL library.

The main steps to create your own application are:

• Modify main.c (see LVGL examples)
• Add any folders and files to extend the functionalities
• Update Dockerfile to add any package
• Modify CMakeLists.txt provided file to ensure all the required files are compiled and linked
• Use the docker scripts provided to build the application for ARM64 architecture

If there is any problem with the output folder generated permissions, modify the permissions:

sudo chown -R $(whoami):$(whoami) lvgl_port_linux/bin

This error can be printed when running the application:

[Warn]	(1382.767, +37)	 lv_display_refr_timer: No draw buffer lv_refr.c:374
[Warn]	(1382.804, +37)	 lv_display_refr_timer: No draw buffer lv_refr.c:374
[Warn]	(1382.841, +37)	 lv_display_refr_timer: No draw buffer lv_refr.c:374
[Warn]	(1382.878, +37)	 lv_display_refr_timer: No draw buffer lv_refr.c:374

To fix the issue find the existing fbdev available:

ls /dev/fb*

Export the variable to match the fbdev name:

export LV_LINUX_FBDEV_DEVICE=/dev/fb0

While running the application, if there is an error about XDG_RUNTIME_DIR, add the following environment variable on the board.

export XDG_RUNTIME_DIR=/run/user/1000

CMake may have troubles with CMakeLists.txt changes with some variables setup. If there is any problem building, try to clean the build folder:

rm -rf lv_port_linux/build-arm64

If you find any issues with the development board feel free to open an Issue in this repository. For LVGL related issues (features, bugs, etc) please use the main lvgl repository.

If you found a bug and found a solution too please send a Pull request. If you are new to Pull requests refer to Our Guide to learn the basics.