Re: Your patch: "recipes-bsp: drop custom deploy location for TF-A binary"
Andrey Zhizhikin
Hello Patrick,
I'll Cc: the meta-freescale ML here, since this a second issue of the same type reported on the TF-A changes. I guess it would be informative for the others to read it in order to avoid duplication. On Wed, Aug 10, 2022 at 12:03 PM Patrick Boettcher <patrick.boettcher@...> wrote: Then I suggest you switch from NXP BSP and move to Mainline one. `meta-imx` implements NXP-specific BSP and is maintained by NXP engineers. This involves also "locking up" versions of all layers in their manifest (the one you used in initial cloning). This in turn implies that if you start to mix-and-match versions of NXP BSP with OSS layer updates - you get those inconsistencies which would need to be solved by person who does this. The process goes in the opposite direction: `meta-imx` releases are merged into `meta-freescale` for those components that are NXP-specific, which allows Mainline BSP to be built as a combination of latest upstream packages and latest NXP vendor parts. This has to be picked up by NXP from `meta-freescale` and integrated into `meta-imx` with consecutive releases. -- Regards, Andrey. |
|