IF573 integration on Jetson Orin Nano (lrd-11.x)

Introduction

This is a comprehensive guide to integrating the Sona IF573 Wi-Fi radio module with the NVIDIA Jetson Orin Nano. This guide is written to assist developers, engineers, and technology enthusiasts in seamlessly incorporating our Sona IF573 Wi-Fi module into their Jetson Orin Nano projects.

Purpose of the Guide

The Jetson Orin Nano, known for its powerful computing capabilities in AI and machine learning applications, becomes even more versatile when combined with the Sona IF573, a high-performance Wi-Fi module. This combination unlocks new possibilities in areas like IoT, autonomous robotics, and smart devices. However, the process of integrating a sophisticated Wi-Fi module with the Jetson platform can present unique challenges, ranging from hardware setup and driver configuration to ensuring optimal performance and stability.

This guide aims to address these challenges in a structured manner. It is designed to provide:

1. Step-by-Step Instructions: Clear, concise, and easy-to-follow steps detailing the hardware setup, ensuring that your Sona IF573 module is correctly connected to the Jetson Orin Nano.

2. Software Configuration: Guidance on configuring the necessary drivers and software on the Jetson Orin Nano to ensure full compatibility with the Sona IF573 module.

3. Optimization Tips: Tips and best practices for optimizing Wi-Fi performance and reliability, crucial for applications requiring robust wireless communication.

4. Troubleshooting Advice: Solutions to common issues that might arise during the installation and operation of the Sona IF573 with the Jetson Orin Nano, saving you time and frustration.

By the end of this guide, you will have a fully functional, high-performance wireless setup on your Jetson Orin Nano.

The Jetson Orin Nano

Proper Installation of the IF573 in the M.2 Slot on the Jetson Development Board

IMPORTANT: The IF573 requires a tri-band antenna, while the Jetson Carrier Board antenna is dual band. Also, an additional antenna must be added for Bluetooth.

• FlexMIMO 6E EFD2471A3S-10MH4L PIFA MHF4L
  https://www.ezurio.com/documentation/datasheet-flexmimo-6e Connects to A0 and A1 (WLAN) ports on the Sona™ IF573 M.2 Module on the DVK.

• FlexPIFA 6E EFB2471A3S-10MH4L PIFA MHF4L
  https://www.ezurio.com/documentation/datasheet-flexpifa-6e Connects to BT (Bluetooth™) port on the Sona™ IF573 M.2 Module on the DVK.

IMPORTANT NOTES:

• The host computer, or build server, must have Ubuntu 18.04 or 20.04

• On the Jetson Orin Nano the SDIO feature is not supported in software. For WiFi/Bluetooth use cases, recommendation is to use PCIe. https://docs.nvidia.com/jetson/archives/r35.4.1/DeveloperGuide/text/SO/JetsonOrinSeries.html#not-supported-features

• The Sona IF573 software version used in this guide is version 11.0.0.138

Toolchain = Aarch64 gcc-9.3-glibc-2.31

• https://docs.nvidia.com/jetson/archives/r35.4.1/DeveloperGuide/text/SO/JetsonOrinSeries.html#not-supported-features

Linux Kernel = Version 5.10.120

• https://docs.nvidia.com/jetson/archives/r35.4.1/DeveloperGuide/text/SO/JetsonOrinSeries.html#kernel

Bootloader Notes

• https://docs.nvidia.com/jetson/archives/r35.4.1/DeveloperGuide/text/SO/JetsonOrinSeries.html#bootloader

Release Notes for Jetpack r35.4.1

• https://docs.nvidia.com/jetson/archives/r35.4.1/ReleaseNotes/Jetson_Linux_Release_Notes_r35.4.1.pdf

1. Register for a NVIDIA developer account

• https://developer.nvidia.com/

2. Install essential host packages (Copy and paste as a whole block)

sudo apt install gawk wget git diffstat unzip texinfo gcc build-essential \
chrpath socat cpio python3 python3-pip python3-pexpect xz-utils debianutils \
iputils-ping python3-git python3-jinja2 libegl1-mesa libsdl1.2-dev python3-subunit \
mesa-common-dev zstd liblz4-tool file locales flex screen libncurses-dev \
perl-modules imagemagick graphviz dvipng latexmk librsvg2-bin texlive-lang-chinese \
texlive-xetex libtest-pod-perl python3-sphinx libssl-dev libelf-dev bison \
libnl-3-dev libdbus-1-dev

3. Install the NVIDIA sdkmanager

Note: sdkmanager can be installed graphically or via the command line. Command line installation information can be found here: https://docs.nvidia.com/sdk-manager/sdkm-command-line-install/index.html

cd ${HOME}/Downloads
wget https://developer.nvidia.com/downloads/sdkmanager/secure/clients/sdkmanager-2.0.0.11402/sdkmanager_2.0.0-11402_amd64.deb
sudo apt install ./sdkmanager_2.0.0-11402_amd64.deb
sdkmanager --cli

Note: If the host machine has a UI you can skip the Command Line Setup and launch sdkmanager from the UI if prefered.

Follow the NVIDIA guide to flash the Jetson with a default image from the sdkmanager.

Important: INitial flashing with the demo image is required to program the EEPROM with the bootloader data.

4. Get the Cross Compilation Toolchain

sudo mkdir /opt/gcc-9.3-glibc-2.31
cd /opt/gcc-9.3-glibc-2.31
sudo wget -O gcc-9.3-glibc-2.31.tar.gz https://developer.nvidia.com/embedded/jetson-linux/bootlin-toolchain-gcc-93
sudo tar -xzvf gcc-9.3-glibc-2.31.tar.gz -C ./

5. Source environment variables.

Note: These environment variables must be sourced each time a new terminal in opened while following this guide. (Copy and paste as a whole block)

export LAIRD_RELEASE=11.0.0.138 ; \
export L4T_JETPACK_VER=JetPack_5.1.2_Linux_JETSON_ORIN_NANO_TARGETS ; \
export L4T_NAME=jetson ; \
export L4T_BRANCH=35 ; \
export L4T_MAJOR=4 ; \
export L4T_MINOR=1 ; \
export L4T_VERSION="${L4T_BRANCH}.${L4T_MAJOR}.${L4T_MINOR}" ; \
export L4T_RELEASE_TAG="${L4T_NAME}_${L4T_VERSION}" ; \
export L4T_KERNEL_DIR=kernel-5.10/ ; \
export L4T_KERNEL_VER=5.10.120 ; \
export L4T_KERNEL_VER_SUFFIX=-tegra ; \
export L4T_TOOLCHAIN_PREFIX=/opt/gcc-9.3-glibc-2.31/bin/aarch64-buildroot-linux-gnu- ; \
export L4T_JETPACK_HOME="${HOME}/nvidia/nvidia_sdk/${L4T_JETPACK_VER}/Linux_for_Tegra" ; \
export L4T_KSRC_DIR="${L4T_JETPACK_HOME}/sources/kernel" ; \
export L4T_KBUILD_DIR_OUT="${L4T_KSRC_DIR}/build" ; \
export L4T_KMODULES_DIR_OUT="${L4T_KSRC_DIR}/modules" ; \
export ARCH=arm64 ; \
export CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX}

6. Get the Kernel Source and Apply NVIDIA Binaries

cd ${L4T_JETPACK_HOME}
./source_sync.sh -t ${L4T_RELEASE_TAG}
sudo ./apply_binaries.sh

7. (Optional) Backup a Copy of Synced Sources

Note: Backing up a fresh copy of the source is useful if the build needs to be restarted. This is optional and not required.

cd ../..
sudo cp -ra ./JetPack_5.1.2_Linux_JETSON_ORIN_NANO_TARGETS ./JetPack_5.1.2_Linux_JETSON_ORIN_NANO_TARGETS.SOURCE_SYNCED

8. Compile Kernel

cd ${L4T_KSRC_DIR}
make -C ${L4T_KERNEL_DIR}/ ARCH=arm64 mrproper
make -C ${L4T_KERNEL_DIR}/ ARCH=arm64 O=${L4T_KBUILD_DIR_OUT} LOCALVERSION=-tegra CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} tegra_defconfig
make -C ${L4T_KERNEL_DIR}/ ARCH=arm64 O=${L4T_KBUILD_DIR_OUT} LOCALVERSION=-tegra CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} menuconfig

Important: In the menuconfig the following items need to be removed:

• deselect Device Drivers -> Network device support -> Wireless LAN
• deselect Device Drivers -> Misc drivers -> Bluedroid_pm driver support
• deselect Networking support -> Bluetooth subsystem support
• deselect Networking support -> Wireless

make -C ${L4T_KERNEL_DIR}/ ARCH=arm64 O=${L4T_KBUILD_DIR_OUT} LOCALVERSION=-tegra CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} -j8
make -C ${L4T_KERNEL_DIR}/ ARCH=arm64 O=${L4T_KBUILD_DIR_OUT} LOCALVERSION=-tegra CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} INSTALL_MOD_PATH=${L4T_KMODULES_DIR_OUT} modules_install

9. Copy kernel to nvidia sdk manager directory

cd ${L4T_KBUILD_DIR_OUT}/arch/arm64/boot
cp -r * ${L4T_JETPACK_HOME}/kernel/

10. Delete old kernel modules in Nvidia SDK Manager directory (kernel)

cd ${L4T_JETPACK_HOME}/rootfs/usr/lib/modules/${L4T_KERNEL_VER}${L4T_KERNEL_VER_SUFFIX}/kernel
sudo rm -rf *

11. Copy kernel modules from build over to Nvidia SDK Manager directory (rootfs)

cd ${L4T_KMODULES_DIR_OUT}
sudo cp -r * ${L4T_JETPACK_HOME}/rootfs/usr
cd ${L4T_JETPACK_HOME}/rootfs/usr/lib/modules
sudo ln -s ${L4T_KERNEL_VER}${L4T_KERNEL_VER_SUFFIX}/ ${L4T_KERNEL_VER}

In the IF573 software package find the backports and the IF573 firmware and copy these archives into $L4T_KSRC_DIR directory.

backports-laird-11.0.0.138.tar.bz2
laird-if573-pcie-firmware-11.0.0.138.tar.bz2
summit_supplicant-src-11.0.0.138.tar.gz

12. Extract the IF573 PCIe firmware to the Nvidia SDK Manager rootfs directory (${L4T_JETPACK_HOME}/rootfs/)

cd $L4T_KSRC_DIR
sudo tar xvf laird-if573-pcie-firmware-${LAIRD_RELEASE}.tar.bz2 -C ${L4T_JETPACK_HOME}/rootfs/

Note: This will extract the following files:

lib/firmware/brcm/CYW55560A1.hcd
lib/firmware/cypress/CYW55560A1_v001.002.087.0225.0065.hcd
lib/firmware/cypress/cyfmac55572-pcie.trxse
lib/firmware/cypress/cyfmac55572-pcie-prod_v18.53.212.28.trxse
lib/firmware/cypress/cyfmac55572-pcie.txt
lib/firmware/cypress/cyfmac55572-if573.txt
lib/firmware/cypress/cyfmac55572-if573_v20231020.clm_blob
lib/firmware/cypress/cyfmac55572-pcie.clm_blob
LICENSE

13. Extract the Backports and the WPA Supplicant

tar xvf backports-laird-${LAIRD_RELEASE}.tar.bz2
tar xvf summit_supplicant-src-${LAIRD_RELEASE}.tar.gz

14. Compile the Backports and the WPA Supplicant

cd $L4T_KSRC_DIR/laird-backport-${LAIRD_RELEASE}
export KLIB_BUILD="${L4T_KBUILD_DIR_OUT}"
export KLIB="${L4T_KMODULES_DIR_OUT}"
make ARCH=arm64 CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} defconfig-lwb
make ARCH=arm64 CROSS_COMPILE=${L4T_TOOLCHAIN_PREFIX} -j8

15. Find the backport modules and tar and extract to Nvidia SDK Manager directory

find . -name *.ko -print0 | tar -cvf laird_modules.tar --null -T -
sudo mkdir ${L4T_JETPACK_HOME}/rootfs/lib/modules/${L4T_KERNEL_VER}/updates
sudo tar xvf laird_modules.tar -C ${L4T_JETPACK_HOME}/rootfs/lib/modules/${L4T_KERNEL_VER}/updates

Note: These are the files that will be extracted:

./compat/compat.ko
./drivers/bluetooth/btbcm.ko
./drivers/bluetooth/btusb.ko
./drivers/bluetooth/hci_uart.ko
./drivers/net/wireless/broadcom/brcm80211/brcmfmac/brcmfmac.ko
./drivers/net/wireless/broadcom/brcm80211/brcmutil/brcmutil.ko
./net/bluetooth/hidp/hidp.ko
./net/bluetooth/bluetooth.ko
./net/bluetooth/bnep/bnep.ko
./net/bluetooth/rfcomm/rfcomm.ko
./net/wireless/cfg80211.ko

16. Create User and Create a SD Card Image

cd ${L4T_JETPACK_HOME}

Optional - Create a username and password

Note: This will create an image with the user ezurio and the password connect.

sudo ./tools/l4t_create_default_user.sh -u ezurio -p connect -a -n jetson-orin --accept-license

Create the SD Card Image

sudo ./tools/jetson-disk-image-creator.sh -o sd-jetson-if573_${LAIRD_RELEASE}.img -b jetson-orin-nano-devkit -d SD

Note: Use Balena Etcher software to flash the image to an SD card.

Download Etcher here: https://etcher.balena.io/#download-etcher

17. Final Setup and Regulatory Domain Configuration on Jetson Nano

Login to the Jetson Nano

Note: If a user and password were not defined in step 16 there is a setup after flashing that is required to login, this is done with the graphical interface or via serial port console. This tutorial assumes the user knows this and how to finish setting up the Jetson Nano.

Option A - Module Parameter Configuration (Recommended for Development Purposes)

The Sona IF573 and Sterling LWBx radios support regulatory domain configuration using a brcmfmac module parameter. The module parameter is called ‘regdomain’, and it takes a country code string. The string must be one of the values listed as a Country Code in the per-radio regulatory release notes for the radio (or ‘ETSI’ for use in EU countries), and can only be used if the radio type in use supports that country. Module parameter configuration is the easiest to implement however it may not be secure from end user modification. Device tree configuration is preferred for a production system and will override the module parameter if both are present.

Important: The driver will fail to load unless the regulatory domain is configured properly.

# Sample contents of /etc/modprobe.d/brcmfmac.conf configured for operation in
#  the United States
options brcmfmac regdomain="US"

Note: The initial list of supported regulatory domains includes US, CA, AU, JP, ETSI

Option B - Device Tree Configuration (When Building a Production System)

The Sona IF573 and Sterling LWBx radios support regulatory domain configuration using the device tree. The configuration must be located in the device tree node corresponding to the Wi-Fi portion of the radio. The node path will vary based on host platform architecture. The node must contain the standard brcmfmac ‘compatible’ property. The regulatory domain is configured with a new property called ‘laird,regdomain’. This property must be set to one of the values listed as a Country Code in the per-radio regulatory release notes for the radio (or ‘ETSI’ for use in EU countries). The string can only be used if the radio type in use supports that country. Device tree configuration is the preferred mechanism because it is secure against end user modification. Device tree configuration will override module parameter configuration if both are present.

# Sample device tree node for a PCIE/UART Sona IF573 module configured for
# operation in the United States
# Note that the compatible string must be "brcm,bcm4329-fmac"
 
&pcie {
    pcie@0 {
        reg = <0x0000 0 0 0 0>;
        #address-cells = <3>;
        #size-cells = <2>;
 
        brcmf_pcie: wifi@1 {
            reg = <0x000000 0 0 0 0>;
            compatible = "brcm,bcm4329-fmac";
 
            laird,regdomain = "US";
        };
    };
};

Run depmod

This will probe all the modules on the system and allow the new modules to be seen.

sudo depmod -a
sudo reboot