Custom Configuration

Device Tree Overlay Configuration

Technexion implement a device-tree overlay method using AVB 3.0.

Example: runtime stage

Please boot up into u-boot prompt, and issue commands as follows example:

Overlay vl10112880 lvds panel
u-boot=> setenv overlay_name 'lvds_10'

Overlay vl215192108 lvds panel
u-boot=> setenv overlay_name 'lvds_21'

Change back original device tree without overlay (HDMI)
u-boot=> setenv overlay_name 'no_overlay'

save your overlay configuration for every boot later, if you just want to boot current setting once, you can ignore this step.
u-boot=> saveenv

boot into Android system
u-boot=> boot

Overlay Table

Example: compile stage

You also can fixed your overlay setting in u-boot header file according your target board such as:

<source>/vendor/nxp-opensource/uboot-imx/include/configs/<your board>_android.h
"overlay_name=<change to what you want>\0"» \

Change Boot Arguments

According to your device, you might need to change properties let your system performance better, list common properties as follows:

• screen rotation
  Because of some DRM-panel driver has clock issue in GKI kernel, so we provide standard hwrotation property or a workaround "swrotation" property in u-boot arguments according to your platform:
  EDM-G-IMX8M-PLUS = hwrotation (0, 90, 180, 270)
  EDM-G-IMX8M-MINI = N/A
  PICO-IMX8M = swrotation(0, 90)
  Note that if you use PICO-IMX8M with Camera and you would like to rotate 90 angle in camera preview mode, you can set swrotation=90 to implement it.
  PICO-IMX8M-MINI = N/A

• cma size = cma

• resolution = displaymode

• WiFi country code for AP mode =wificountrycode

• display density =lcd_density

About CMA size, it's up to your DRAM size for each platforms, please modify CMA size as follows table:

Example: runtime stage

EDM-G-IMX8M-PLUS

u-boot=> setenv bootargs console=ttymxc1,115200 init=/init androidboot.console=ttymxc1 androidboot.hardware=nxp androidboot.hwrotation=0 firmware_class.path=/vendor/firmware loop.max_part=7 transparent_hugepage=never swiotlb=65536 androidboot.vendor.sysrq=1 androidboot.lcd_density=240 androidboot.primary_display=imx-drm androidboot.wificountrycode=TW androidboot.usb.debugging=1 cma=800M@0x400M-0xb80M quiet
u-boot=> saveenv
u-boot=> boot

EDM-G-IMX8M-MINI

u-boot=> setenv bootargs console=ttymxc1,115200 init=/init androidboot.console=ttymxc1 androidboot.hardware=nxp firmware_class.path=/vendor/firmware loop.max_part=7 transparent_hugepage=never androidboot.vendor.sysrq=1 androidboot.lcd_density=240 androidboot.primary_display=imx-drm androidboot.wificountrycode=TW androidboot.usb.debugging=1 cma=480M@0x400M-0xb80M quiet
u-boot=> saveenv
u-boot=> boot

EDM-G-IMX8M-NANO

u-boot=> setenv bootargs console=ttymxc1,115200 init=/init androidboot.console=ttymxc1 androidboot.hardware=nxp firmware_class.path=/vendor/firmware loop.max_part=7 transparent_hugepage=never androidboot.vendor.sysrq=1 androidboot.lcd_density=240 androidboot.primary_display=imx-drm androidboot.wificountrycode=TW androidboot.usb.debugging=1 cma=800M@0x400M-0xb80M quiet
u-boot=> saveenv
u-boot=> boot

PICO-IMX8M

u-boot=> setenv bootargs console=ttymxc0,115200 init=/init androidboot.console=ttymxc0 androidboot.hardware=nxp androidboot.hwrotation=0 firmware_class.path=/vendor/firmware loop.max_part=7 androidboot.fbTileSupport=enable transparent_hugepage=never androidboot.vendor.sysrq=1 androidboot.lcd_density=240 androidboot.primary_display=imx-dcss androidboot.displaymode=1080p androidboot.wificountrycode=TW androidboot.usb.debugging=1 cma=736M quiet
u-boot=> saveenv
u-boot=> boot

PICO-IMX8M-MINI

u-boot=> setenv bootargs console=ttymxc1,115200 init=/init androidboot.console=ttymxc1 androidboot.hardware=nxp androidboot.swrotation=90 firmware_class.path=/vendor/firmware loop.max_part=7  transparent_hugepage=never androidboot.vendor.sysrq=1 androidboot.lcd_density=240 androidboot.primary_display=imx-drm androidboot.wificountrycode=TW androidboot.usb.debugging=1 cma=400M@0x400M-0xb80M quiet
u-boot=> saveenv
u-boot=> boot

Example: compile stage

You also can fixed your overlay setting in u-boot header file according your target board such as:

<source>/vendor/nxp-opensource/uboot-imx/include/configs/<your board>_android.h
"bootargs=<change to what you want>\0"» \

Enableing NFC function

Users can enable NFC function in < source >/cookers/env.bash (default is disabled except for EDM-G-IMX8M-PLUS), then re-compile again:

export NFC_ACTIVE=false
to
export NFC_ACTIVE=true

Enableing 5GNR function

Support platform: EDM-G-IMX8M-PLUS

Step 1: Prepare SIM8202 prebuilt Radio Interface Layer (RIL) libraries

This BSP does support SIMCOM 5GNR module - SIM8202G-M2, but because of the license restriction, please contact TechNexion FAE or Sales to get licensed RIL relate files, default is disabled.

Contact Window: sales@technexion.com

After getting the RIL files, please decompress the tarball and put modem folder into specific device folder as follows:

$ cp -rv modem <source folder>/device/nxp/edm_g_imx8mp

Step 2. Enable SIM8202 configuration

Users need enable 5gnr function in < source >/cookers/env.bash (default is disabled, note that modify top SIM8202_MODEM_ACTIVE parameter is enough), then re-compile again:

export SIM8202_MODEM_ACTIVE=false
to
export SIM8202_MODEM_ACTIVE=true

LIBGPIOD JNI APIs

Technexion provide a demo app about libgpiod JNI Test, specific source code as following:

• source code

Users can implement own GUI using our INPUT/OUTPUT APIs

Setting GPIO as output with specific value:
public native String  setGpioInfo(int gpiobank,int gpioline, int value)

Setting GPIO as input and get a value:
public native String  getGpioInfo(int gpiobank,int gpioline);

New OTA system: systemupgrade

Our Android 11 does support full update and incremental update, please contact our window to get detail instruction:
sales@technexion.com

TensorFlow-Lite for Machine Learning

EDM-G-IMX8MP has a NPU (Neural Processing Unit) inside the SoC for machine learning applications, it has a very good performance using TFL (tensorflow-lite), Technexion provide two simple TFL apps let users testing as follows steps:
Note that other platforms

Image classification using CPU/NPU

$ wget https://ftp.technexion.com/development_resources/NXP/android/11.0/machine-learning_demo/TFL-ImageClassify.zip

$ unzip TFL-ImageClassify.zip

Speech commands using CPU

$ wget https://ftp.technexion.com/development_resources/NXP/android/11.0/machine-learning_demo/TFL-SpeechCommand.zip
$ unzip TFL-SpeechCommand.zip

Quick install to Android image

$ adb install TFL-ImageClassify.apk
$ adb install TFL-SpeechCommand.apk

Build-in to Technexion Android BSP

$ mkdir -p <source code>/packages/apps/TFL-ImageClassify
$ mv TFL-ImageClassify.apk Android.mk <source code>/packages/apps/TFL-ImageClassify
$ cd <source code>/packages/apps/TFL-ImageClassify
$ mm
$ cd -
$ cook -j4

$ mkdir -p <source code>/packages/apps/TFL-SpeechCommand
$ mv TFL-ImageClassify.apk Android.mk <source code>/packages/apps/TFL-SpeechCommand
$ cd <source code>/packages/apps/TFL-SpeechCommand
$ mm
$ cd -
$ cook -j4