Add support for more boards #125
Conversation
Rewrite flash recipe to use a data-driven approach to board definitions instead of duplicating logic for each board. Signed-off-by: Loïc Minier <[email protected]>
Instead of using `unzip -j` which will ignore the directory structure inside zip archives, manually move files out of the toplevel directories. This avoids errors if the same filename is used multiple times, for instance prog_firehose_ddr.elf might be present multiple times in some boot binaries zip archives (e.g. for the sail/ subdirectory). Signed-off-by: Loïc Minier <[email protected]>
Also known as QCS9100 Ride Rev3 EVK Signed-off-by: Loïc Minier <[email protected]>
Also known as QCS8300 Ride SX EVK Signed-off-by: Loïc Minier <[email protected]>
|
Tested for RB1 and RB3Gen2 by downloading CI artifacts dtb.tar.gz and rb1-boot.img, running this before and after the changes: then compared the resulting flash subdirectories, and these were identical except for UUID changes to the GPT partition tables and the dtb.bin image. |
Test jobs for commit ac618ac |
|
I've submitted the build to LAVA for IQ-8, I took the rb3 config but had to remove the cambridge-lab tag as we only have one board in Hyderabad. Unfortunately, flashing times out: |
Signed-off-by: Loïc Minier <[email protected]>
Test jobs for commit 0563494 |
|
I submitted the last CI build for this branch ( I'd still want to merge this as to have a starting point for debugging. The next thing to research is whether the mainline kernel boots. |
ricardosalveti
left a comment
ricardosalveti
left a comment
There was a problem hiding this comment.
ride-sx does not boot but I was able to check that the content is correct, it might just be missing kernel configs for it.
|
Did a manual boot of these boards using mainline kernel in LAVA: both worked fine with 6.17-rc5 |
2 participants
Changes to debos recipes to have a more scalable board support approach and add
support for a couple more boards:
Commits: