*** Please note that this reply is generated by LLM automatically ***
The question is about understanding the flash process on a Jetson Orin NX device using the NVIDIA SDK Manager. Specifically, the user is trying to identify the script that calls the tegraflash.py command to read the EEPROM and reboot to recovery mode.
After analyzing the provided log snippet, it appears that the tegraflash.py command is called by the NVL4TFLASHJETSONLINUXCOMP.sh script, which is executed by the nvsdkmanager_flash.sh tool. This script is part of the NVIDIA SDK Manager and is responsible for flashing the Jetson device.
Here’s a breakdown of the relevant log lines:
Line 2367: info NVL4TFLASHJETSONLINUXCOMP:./exec_command: /home/vuh81/hcn/nvsdkm/replay/scripts/JetPack6.2.1/Linux/NVL4TFLASHJETSONLINUXCOMP.sh
This line indicates that the NVL4TFLASHJETSONLINUXCOMP.sh script is being executed.
Line 2154: info: Reading EEPROM: /home/vuh81/hcn/nvidia/nvidia_sdk/JetPack6.2.1/Linux/JETSON_ORIN_NX_TARGETS/Linux_for_Tegra/bootloader/tegraflash.py ...
This line shows that the tegraflash.py command is being called to read the EEPROM.
Comparing this to the provided context, we can see that the nvsdkmanager_flash.sh tool is a flashing interface script for the SDK Manager, and it is likely responsible for calling the NVL4TFLASHJETSONLINUXCOMP.sh script, which in turn calls the tegraflash.py command.
To summarize:
The nvsdkmanager_flash.sh tool is the top-level script that initiates the flash process.
The NVL4TFLASHJETSONLINUXCOMP.sh script is called by nvsdkmanager_flash.sh and is responsible for executing the tegraflash.py command to read the EEPROM and reboot to recovery mode.
If you’re interested in modifying the flash process or skipping the EEPROM read, you may want to explore the SKIP_EEPROM_CHECK=1 option, which is mentioned in the context as a way to skip the EEPROM read during flashing. However, this would require further investigation and testing to ensure it works as intended.
*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***
Use case is just a generic flashing sequence in SDK Manager: open it, choose pre-config and input my username & password, select which to download and install (Host side and Device side), then click Start, and after flashing complete I retrive the log.
If it was called by nvautoflash.sh, then how can it be called before nvsdkmanager.sh? Because looking into the code I thought that nvsdkmanager.sh is the one that calls nvautoflash.sh (by using --nv-auto-config flag)
I haven’t tried flashing by initrd flash script yet.
If you want to flash through command, we would suggest using initrd flash script.
Please refer to <Linux_for_Tegra>/tools/kernel_flash/README_initrd_flash.txt for details.
Let me also share an example flash command as following: