On 2/21/25 3:45 PM, Adrian Freihofer wrote:
[...]
Having said all of that, notice that if the decision was really to
switch to interpretation (1) then other changes would be required
besides those made to concat_dtb(). In particular, I imagine
fitimage_emit_section_config() would require changes to avoid
emitting
the signature node for the configurations.
Hope this clarifies things...
Yes, from my point of view that is the case. I came to exactly the same
conclusion. The use of FIT_SIGN_INDIVIDUAL is strange, but not wrong.
It means that it's definitely not something that can be changed on an
LTS branch. Although I really like Marek's patch, we need to revert
that part of it.
That would break the use of signed fitImages with embedded env script,
which does not work either.
I've posted the following patch instead, you should all be on CC:
[PATCH] u-boot: kernel-fitimage: Restore FIT_SIGN_INDIVIDUAL="1" behavior
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#211816):
https://lists.openembedded.org/g/openembedded-core/message/211816
Mute This Topic: https://lists.openembedded.org/mt/111218371/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
