I want to create a persistent partition after APP in eMMC. If the partition does not exist, I want it to flash an empty ext4 image of the correct size. If it already exists, I want to leave it alone in the flashing procedure.
Is this something I would do from jetson-tk1.conf/flash.sh/gnu_linux_fastboot_emmc_full.cfg, or should I just shrink the partition and create the file system on target after the image starts up?
I managed to add my partition quite easily.
[partition]
name=TST
id=42
type=data
allocation_policy=sequential
filesystem_type=basic
size=536870912
file_system_attribute=0
partition_attribute=0
allocation_attribute=8
percent_reserved=0
#filename=system.img
But I don’t understand how this partitioning works. I’ve lowered the rootfs size by 2GB, but if I try to set the TST partition to 1GB I get an error message saying there isn’t room for the UDA partition. Any documentation on how this partitioning works? How to build a valid config file, and what partitions are needed when running l4t?
If total size of all the partition exceeds your emmc size, that’s what the complaint is. Also noticed that your partition ID is 42. If you check id of each partition, they are in sequences and start from 2 and end with 20/21 of UDA/GPT. UDA and GPT pairs have to be the last 2 partition, user data area and system global partition table. For your case, you could utilize TP1, TP2 or TP3 partition. Those partitions are temporary place holder and can be used for your own purpose. Again, be aware of total size.
We will try to put these information in elinux.org. Thanks for your feedback.
Did not see your reply until now. Will try what you suggested. Thanks.
That worked. Thanks :)