I am trying to understand the mender partition layout and the storage of the uboot environment better when using eMMC storage with u-boot on an i.MX8m.
Our SOM vendor Compulab as well as Toradex (https://developer.toradex.com/knowledge-base/emmc-linux#Storage_and_Partitioning) seem to recommend putting uboot into the boot partitions (/dev/mmcblk0boot0) on the eMMC instead of as a part of the image which our mender config does.
This leads to a couple of questions:
- Is it recommended with Mender to put uboot into the boot partition on the eMMC or to include it in the main storage space?
- What would need to be changed in our config and programming process to make that change? I am guessing this would require changes to some of these variables:
as well as having a separate dd step to copy the uboot into the boot partition.
- Does moving where the uboot is stored change where the uboot env is at?
For the i.MX platform I have seen some examples that have multiple ENV_STORAGE_DEVICE_OFFSETs. Is there documentation that describes how that is supposed to be set? Should we always have multiple environments? Are these offsets arbitrary or where do they come from?
- I am working with the Dunfell yocto release so are there any specific issues with this approach that I should be aware of when using the libubootenv instead of uboot_fw_utils?