I am not quite sure what you are referring to here resp. no further setenv commands are required for an Ethernet update as outlined above.
The background of my question is that I performed some cleanup of not required u-boot env variables which are compiled into u-boot as default env. (The reason is that we have enough space in the future and do not reach the env size limit which would require re-partitioning of the emmc.)
The idea is to setup your development network as outlined in the following article on our developer website. Then you run our update.sh script with the -o argument pointing to your tftpboot folder. Now subsequent flashing over Ethernet e.g. by using TFTP should work out-of-the-box either as fallback when doing run setupdate or explicitly doing run setethupdate either one followed by run update.
As I tested the flashing as pass I should be fine.
There is no such thing as flashing over the recovery mode. The recovery mode is USB only allowing you to execute something like e.g. U-Boot on the target but it does not flash anything per se. That said of course you may modify it to automatically flash your desired variant as well.
Does that mean that if I am able to mount the image partitions with “ums 0 mmc 0” (ums works over usb as well, right?) the recovery mode is working with my u-boot as well?
That is basically a single entry point allowing for more advanced scripting options.
Ok. The point which is not clear for me is whether the recovery mode over usb requires additional usb functionality in addition in comparison to “ums 0 mmc 0”?
I am not aware of any flash_mmc.src resp. that one should nowadays be called flash_blk.src. And BTW not the .src ones but rather their mkimage generated .img files are what is ultimately used during flashing.
Ok. (We just want to keep the sources and build scripts in version control.)