Re: Aw: Re: Aw: [PATCH v3 1/2] arm64: dts: mt7986: add dtbs with applied overlays for bpi-r3

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 06/11/24 19:49, frank-w@public-files.de ha scritto:
> Hi
> 
> any new state on this??
> 
> https://patchwork.kernel.org/project/linux-mediatek/patch/20240608080530.9436-2-linux@fw-web.de/
> 
> regards Frank

I had a look at this one - and this is the situation:

  1. Your bootloader supports loading DTBO, so you can indeed use DTBOs
  2. Validation of the DTSO can still be done during kernel build without adding
     all of those possible X+Y+Z pre-joined DTBs
  3. What if your hardware had more than 20 possible configurations?
     Would you write 20 different Makefile entries? "sd+nand+nor",
     "sd+nand-withoutnor", "emmc+nand+nor", "emmc+nand-withoutnor",
     "emmc+sd+nor", "emmc+sd-withoutnor", "ufs+emmc", "ufs+emmc+sd",
     "ufs+sd+nor", "ufs+emmc-withoutnor", "ufs+sd-withoutnor", ......

Looks messy and unfeasible.

However, this is not a *global* no: I'm happy that your bootloader does support
loading DTBOs and, as far as I remember, even uses straps to vary the DTB(o) to
actually load - which is something proper and great... so it's a *no* for you,
but more than just a no, this is "why are you treating your proper bootflow
like it was a broken one?!?!" :-)

If anyone finds themselves in a situation in which there's no way to update a
bootloader (and that unfortunately happens more often than anyone would like
to see...) and in which the only way to apply DTBOs is to pre-overlay them
during the kernel build, then that's fine and I would (if nice and clean)
accept that.

But again, you're not in this kind of situation - and you're lucky that you're
dealing with a fully open device with a proper bootloader and bootflow: don't
ruin it like that!
Instead, if necessary, update the userspace tools that you're using to deal with
multiple DTBOs during system upgrades: that's the right thing to do at this point.

Cheers,
Angelo