18:45 <austriancoder> NishanthMenon: Imagine the following situation: am642 SoC with secure boot.. emmc boot configured.. should it be possible to have an a/b U-Boot mechanism by using boot0/1 partitions?

19:26 <NishanthMenon> austriancoder: what i did for beagle ai64 was only put tiboot3 in boot0 partition. that allowed me flexibility to scale u-boot, optee etc also enabled a/b update capability. tiboot3.bin a/b update - i think we might have a weakness - might be good to ask in e2e.ti.com to confirm an official answer

19:31 <austriancoder> NishanthMenon: would it be possible to switch boot0 -> boot1 if we fail to use U-Boot from boot0?

19:32 <NishanthMenon> austriancoder: if you are talking of R5 u-boot SPL (aka tiboot3.bin content) - I am not sure. but R5 SPL itself, i think might be standard for supporting a/b update scheme..

19:33 <austriancoder> the whole bootflow is quite complex :)

19:34 <austriancoder> I am not a big fan of the e2e forum but I will als there too.

19:34 <austriancoder> I was in hope that would be some best practices from TI.

22:25 <NishanthMenon> austriancoder: yeah - the complexity feedback is being pulled in for the next arch iteration. unfortunately k3 had very strong constraints of need for flexibility, family of device compatibility and orthogonal usecases to support. :(