pgils / docker-jetson-flash Goto Github PK

unpack zipped images (as downloaded from BalenaOS) automatically when passed with the -f flag

With recent L4T 32.5 releases the following error occurs when flashing a jetson-xavier-nx-devkit-emmc:

l4t_sign_image.sh: Generate header for Image_sigheader.encrypt
l4t_sign_image.sh: chip 0x19: add 0x20bf808 to offset  0x8 in sig file
/tmp/Linux_for_Tegra/l4t_sign_image.sh: line 95: xxd: command not found
failed.

This is not observer when flashing using persistent data (-p) from a L4T 32.4 flashing.

• if -o is specified this should be used (relative path should be allowed)
• default: use named docker volume

Currently the tool doesn't run on stock MacOS Big Sur. At a minimum it requires updating bash through homebrew and supporting BSD commond line tools such as readline.

Tags created and pushed by the bump-version workflow do not trigger the build workflow.

some tags are missing from registry (v0.38.0, v0.38.1) when checking out tag v0.38.2 djetson-flash tries to pull tag 0.38.100

➜  docker-jetson-flash git:(v0.38.2) ./djetson-flash
Unable to find image 'ghcr.io/pgils/docker-jetson-flash:v0.38.100' locally

Do we need to set a rootfs size for default BalenaOS images? If we do, don't we end up in trouble when updating BalenaOS through BalenaCloud?