That is because the content needs to be signed. Invalid signatures are not allowed. The unit isn’t bricked, it can still be flashed with a signed tree. Normally this occurs when using the flash tools. Flash tools can be used to create a signed version of a tree, and then dd of the signed tree will work. It is usually easier to just flash without the intermediate signing and dd.
So where is this signing documented.
You have not documented the R32.3.1 procedure enough to know where to put the dtb before I use the flash.sh command.
Let alone the correct name that flash.sh looks for.
The exact file depends on several revisions. Someone else may know the specifics for your case, but you will likely need to post any revisions you have for the carrier board. The reason for flashing once, with a log, is to let the software tell you itself what it is using (which never fails). The development documentation does mention this, but is a bit too generic at times.
Please document the signing procedure, is this something that can be run command line on the actual nvidia and then do the “dd”.
Specifically, this is for creating a signed file without actually flashing:
sudo ./flash.sh <b>--no-flash</b> -u <rsa_priv>.pem -v <sbk>.key jetson-tx2 mmcblk0p1
I have no way to tell you the specific file for your board.